WO2015116834A1

WO2015116834A1 - Compositions and methods for blood biomarker analysis for predicting psychosis risk in persons with attenuated psychosis risk syndrome

Info

Publication number: WO2015116834A1
Application number: PCT/US2015/013555
Authority: WO
Inventors: Clark Jeffries; Diana PERKINS; Elaine WALKER; Tyrone CANNON; Thomas H. MCGLASHAN; Scott W. Woods; Barbara CORNBLATT; Larry SEIDMAN; Kristin CADENHEAD; Ming TSUANG; David MATHALON; Carrie BEARDEN; Jean ADDINGTON
Original assignee: The University Of North Carolina At Chapel Hill; The Regents Of The University Of California; Yale University; Emory University; University Of Calgary; The Feinstein Institute For Medical Research; The Beth Israel Deaconess Medical Center, Inc.
Priority date: 2014-01-29
Filing date: 2015-01-29
Publication date: 2015-08-06

Abstract

Provided herein are methods for identifying a risk of developing a psychotic disorder in a subject. In some embodiments, the methods include isolating serum and/or plasma from a subject who meets clinical criteria for elevated risk of psychosis; quantifying levels of expression in the serum and/or plasma for a plurality of gene products; combining the quantified levels of expression to create a summary measure of expression; and comparing the summary measure of expression for the subject to one or more standards; wherein the comprising step identifies a risk of developing a psychotic disorder in the subject. Also provided are methods for identifying whether or not a subject is developing a psychotic disorder, kits for use in performing the disclosed methods, and non-transitory computer readable medium having stored thereon computer executable instructions that can be used to perform the disclosed methods.

Description

DESCRIPTION

COMPOSITIONS AND METHODS FOR BLOOD BIOMARKER ANALYSIS FOR PREDICTING PSYCHOSIS RISK IN PERSONS WITH ATTENUATED PSYCHOSIS RISK SYNDROME

CROSS REFERENCE TO RELATED APPLICATION

The presently disclosed subject matter is based on and claims the benefit of U.S. Provisional Patent Application Serial No. 61/932,881, filed anuary 29, 2014; the disclosure of which is incorporated herein by reference in its entirety.

GOVERNMENT INTEREST

This invention was made with United States government support under Grant U01 MH082004 awarded by National Institutes of Health of the United States. The United States government has certain rights in the invention.

TECHNICAL FIELD

The presently disclosed subject matter relates to methods for identifying a risk of developing a psychotic disorder in a subject. Also provided are methods for identifying whether or not a subject is developing a psychotic disorder.

BACKGROUND

Psychotic disorders, including schizophrenia, schizoaffective disorder, schizophreniform disorder, brief psychotic disorder, psychotic disorder not otherwise specified, mania with psychotic features, delusional disorder, and major depression with psychotic features, are a family of brain disorders characterized by the presence of psychotic symptoms that may include hallucinations, delusions, and disorganized thought process and/or behaviors. The symptoms significantly interfere with social and vocational function. The majority of psychotic disorders emerge during adolescence and early adulthood. About 1% of the general population develops a psychotic disorder, typically in late adolescence or early adulthood.

It is well established that early intervention is associated with better clinical outcomes in persons with schizophrenia (Perkins et al, 2005), raising the hope that treatment during the "prodromal" phase of illness could prevent the development of a psychotic disorder and thus reduce the risk of an individual developing chronic symptoms and disability. To that end, substantial progress has been made in establishing clinical criteria to identify persons at elevated risk for the development of psychosis, with about 20-25% developing psychosis within a year and 30-35% within two years (Cannon et al, 2008; Woods et al, 2009; Fusar-Poli et al, 2012). These clinical criteria emphasize the presence of attenuated (less than psychotic severity) positive symptoms plus subjective experience of distress and/or impact of symptoms on function. An "Attenuated Psychosis Syndrome" (APS) that defines a high risk syndrome for future development of a psychotic disorder was considered for inclusion in the most recent Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-V). However, the expert panel decided to place the syndrome in the DSM-V Appendix due to concerns that specificity would be low (due to imprecision of diagnosis relative to anxiety and mood disorders) and that persons meeting syndrome criteria actually progress to develop a psychotic disorder (Carpenter & Tandon, 2013; Tsuang et al, 2013). This decision reinforces the need for further research to better define a high risk state for psychosis that would indicate preventative interventions are warranted.

In the practice of medicine, blood tests are routinely used to aid diagnosis, to provide prognosis, and to plan and refine treatment of many disorders such as, but not limited to diabetes (see e.g., Abbasi et al, 2012; Lee & Colagiuri, 2013) and cardiovascular diseases (see e.g., Ahluwalia et al, 2013; Bleakley et al, 2013; Goh et al, 2013). It is important to appreciate that in general, biomarker tests are most useful in persons already determined to be at elevated risk for a disease, typically based on clinical criteria (e.g., obesity as a clinical risk factor for diabetes and cardiovascular disease). This is related to the fact that the proportion of patients who test positive and who actually get the disease (i.e., the positive predictive value; PPV) varies dramatically depending on the actual rate of disease in the tested population. If the rate of disease is low, for example if only 1 % of tested persons are actually at risk for the disease, then even a test with high sensitivity and specificity (-0.8) will have 20 times as many false positives as true positives (see Figure 1).

Thus, to minimize the risk of embarking on an inappropriate treatment program, any biomarker test with less than perfect specificity should ideally be employed in a relatively high risk population. Persons meeting clinical criteria for a high risk of psychosis have about a 30-35% risk of fully developed psychosis within two years (Fusar-Poli et al, 2012), and thus a biomarker test with reasonable sensitivity and specificity (-0.80) can achieve PPV of about 0.63; equivalently, about two-thirds of persons identified by such a test as at risk would truly be on a trajectory to develop psychosis. Of perhaps equal significance is that such a diagnostic test would have a negative predictive ability of -0.90, greatly improving a clinician's confidence in predicting who is likely not to develop psychosis. For this reason it is conceivable that blood assays or other biomarkers with good sensitivity and specificity could have clinical utility in enhancing prediction of psychotic affliction, identifying among patients presenting with "prodromal" signs and symptoms those persons where concern for psychosis is greatest as well as those for whom concern for psychosis is actually quite low.

There are numerous studies that evaluate peripheral blood biomarkers in schizophrenia, and at least one study that evaluated peripheral blood biomarkers in persons prior to the onset of psychosis. Most studies have examined a small number of biomarkers (Miller et al, 2011) comparing persons with schizophrenia, on or off medications, to unaffected persons. A series of studies has been done using a multiplex platform in persons with schizophrenia compared to unaffected persons, and a small number of "prodromal" subjects (Tsang et al, 2006; Craddock et al, 2008; Schwarz et al, 2010; Steiner et al, 2010; Chan et al, 2011; Izmailov et al, 2011; Schwarz et al, 2011; Schwarz et al, 2012a; Schwarz et al, 2012b; Schwarz et al, 2012c; Tomasik et al, 2012; Herberth et al, 2013). These studies have resulted in several patents and/or patent applications for inventions based on levels of blood analytes that purport to diagnose schizophrenia or predisposition thereto. See e.g., U.S. Patent No. 8,492,418 to Woods; PCT International Patent Application Publication Nos. WO 2003/025224 of Bahn; WO 2013/186562 of Bahn et al: U.S. Patent Application Publication No. US 2014/0200151 of Bahn & Schwarz.

SUMMARY

This Summary lists several embodiments of the presently disclosed subject matter and in many cases lists variations and permutations of these embodiments. This Summary is merely exemplary of the numerous and varied embodiments. Mention of one or more representative features of a given embodiment is likewise exemplary. Such an embodiment can typically exist with or without the feature(s) mentioned; likewise, those features can be applied to other embodiments of the presently disclosed subject matter, whether listed in this Summary or not. To avoid excessive repetition, this Summary does not list or suggest all possible combinations of such features. In some embodiments, the presently disclosed subject matter provides a panel of analyte assays (in some embodiments, blood analyte assays) that when used in combination create an index that increases the clinical certainty of psychosis risk in subjects at a substantially higher risk than that of the general population for the development of a psychotic disorder. The panel of analytes includes in some embodiments a summary measure of the levels of malondialdehyde-modified low density lipoprotein (CAS Registry No. 542-78-9), thyroid stimulating hormone (SWISS-PROT Accession No. P01222), Interleukin-IB (SWISS-PROT Accession No. P01584), matrix metalloproteinase-7 (SWISS-PROT Accession No. P09237), and immunoglobulin E (SWISS-PROT Accession No. P01854). In some embodiments, the levels of additional analytes in a sample isolated from a subject (in some embodiments, a blood sample) can be added to this core set of analytes to further increase the sensitivity and specificity of the assay. In some embodiments, the additional analytes include but are not limited to one or more of growth hormone, KIT ligand, interleukin-8, apolipoprotein D, mucin- 16, Factor 7, chemokine (c-c motif) ligand 2, resistin, Cortisol, chemokine (c-c motif) ligand 8, alpha-2-macroglobulin, transthyretin, uromodulin, beta-2 transferrin, prostaglandin D synthase (beta trace-protein), adrenocorticotropin releasing hormone, and insulin-like growth factor.

Subjects meeting high risk criteria for psychosis are frequently prescribed medications, such as but not limited to antipsychotics and antidepressants, which can alter the expression of analytes in the blood and/or serum of subjects. For subjects meeting high risk criteria that are not being treated with antipsychotics, one or more of N-(alpha)- acetyltransferase 15, ferritin, and alpha 1-antiichymotrypsin can in some embodiments be added to the panel to increase sensitivity and specificity. For subjects meeting clinical high risk criteria that are not being treated with antidepressants, in some embodiments chromogranin A and/or endothelin 1 can be added to the panel of core analytes disclosed herein. Analytes can also be expressed differently in males than in females. For males, one or more of interleukin-15, chemokine (c-c motif) ligand 11, and apolipoprotein A2 can in some embodiments be added to the core panel. For females, one or more of calbindin 1, transforming growth factor beta 3, macrophage migration inhibitory factor, pappalysin-1, and testosterone can in some embodiments be added to the core panel.

In some embodiments, the presently disclosed subject matter relates to use of an analyte biomarker panel (in some embodiments, a blood analyte biomarker panel) at two time points, where a change in level of the analyte summary measure and/or of individual analytes can increase the clinical certainty of psychosis risk.

In some embodiments, the presently disclosed subject matter provides using the analyte biomarker panel (in some embodiments, a blood analyte biomarker panel) disclosed herein in combination with other indicators of psychosis risk, including but not limited to premorbid function; symptom severity; social and vocational function; volume or change in volume of brain gray matter, white matter, and/or ventricular volumes; measures of blood brain barrier permeability; electroencephalogram (EEG) measures or change in measures of brain function; brain imaging measures or change in measures of brain function; and/or salivary Cortisol and/or salivary DHEA levels to increase the clinical certainty of psychosis risk in persons meeting high risk criteria for psychosis.

Accordingly, the presently disclosed subject matter provides in some embodiments, methods for identifying a risk of developing a psychotic disorder in a subject. In some embodiments, the presently disclosed methods comprise isolating and/or providing a biological sample comprising serum, plasma, urine, and/or saliva from a subject who meets clinical criteria for elevated risk of psychosis; quantifying a level of expression in the biological sample of a plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample, wherein the gene products or other endogenous or exogenous substances, or the mixture thereof, are selected from the group consisting of malodialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E, combining the quantified levels of expression from step (b) to create a summary measure of expression for the subject; and comparing the summary measure of expression for the subject to one or more standards. In some embodiments, the one or more standards are selected from the group consisting of summary measures of expression in subjects with elevated risk for psychosis who did not develop psychosis; summary measures of expression in subjects with elevated risk for psychosis who did develop psychosis; and summary measures of expression in unaffected subjects; wherein the comprising step identifies a risk of developing a psychotic disorder in the subject. In some embodiments, the summary measure of expression for the subject is compared to both summary measures of expression in subjects with elevated risk for psychosis who did not develop psychosis and summary measures of expression in subjects with elevated risk for psychosis who did develop psychosis in order to determine whether the summary measures of expression for the subject in question more closely approximates (a) the summary measures of expression in subjects with elevated risk for psychosis who did not develop psychosis of (b) the summary measures of expression in subjects with elevated risk for psychosis who did develop psychosis in order to classify the subject in question as being more likely to develop psychosis or not develop psychosis.

The presently disclosed subject matter also provides in some embodiments, methods for identifying whether or not a subject is developing a psychotic disorder. In some embodiments, the presently disclosed methods comprise isolating and/or providing a biological sample comprising serum, plasma, urine, and/or saliva from a subject who meets one or more clinical criteria for elevated risk of psychosis; quantifying a level of expression in the biological sample of a plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample, wherein the gene products or other endogenous or exogenous substances are malodialdehyde-modified low density lipoprotein, thyroid stimulating hormone, mterleukin-lB, matrix metalloproteinase 7, and immunoglobulin E; combining the quantified levels of expression to create a summary measure of expression for the subject; and comparing the summary measure of expression for the subject to one or more standards; repeating the isolating, quantifying, combining, and comparing steps at a second, later time point; and assessing whether the result of the comparing step as determined using the biological sample isolated at the second time point as compared to the result of the comparing step as determined using the biological sample isolated at the first time point is indicative of the subject developing a psychotic disorder.

In some embodiments of the presently disclosed methods, the isolating, quantifying, combining, and comparing steps are performed at least two different time points, and the results of the comparing step from the at least two different time points is indicative of the subject developing a psychotic disorder.

In some embodiments of the presently disclosed methods, the one or more standards are selected from the group consisting of summary measures of expression in subjects with elevated risk for psychosis who did not develop psychosis; summary measures of expression in subjects with elevated risk for psychosis who did develop psychosis; summary measures of expression in unaffected subjects; or any combination thereof. In some embodiments of the presently disclosed methods, at least one of the one or more standards comprises a summary measure that is normalized with respect to healthy subjects of similar age, sex, and/or ancestry as the subject.

In some embodiments of the presently disclosed methods, the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample further comprises one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, interleukin-8, apolipoproptein D, chemokine (c-c motif) ligand 8, factor 7, Cortisol, resistin, alpha-2- macroglobulin, mucin- 16, chemokine (c-c motif) ligand 2. In some embodiments of the presently disclosed methods, the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in serum, plasma, urine and/or saliva further comprises one or more of beta 2 transferrin, prostaglandin D synthase (beta trace protein), adrenocorticotropin releasing hormone, and insulin-like growth factor.

In some embodiments of the presently disclosed methods, the subject is not being treated with an antidepressant, and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample further comprises chromogranin A, endothelin 1, or both.

In some embodiments of the presently disclosed methods, the subject is not being treated with an antipsychotic, and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample further comprises one or more of N-(alpha)-acetyltransferase 15, ferritin, and alpha 1- anti-chymotrypsin.

In some embodiments of the presently disclosed methods, the subject is a female and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found the biological sample further comprises one or more of calbindin 1, transforming growth factor beta, and cytokine (c-c motif) ligand 18.

In some embodiments of the presently disclosed methods, the subject is a male and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample further comprises one or more of interleukin 15 and chemokine (c-c motif) ligand 11.

In some embodiments of the presently disclosed methods, the subject has or is at risk for developing Attenuated Psychosis Syndrome (APS; see the Diagnostic and Statistical Manual of Mental Disorders, Fifth ed. (DSM - 5)). In some embodiments, the presently disclosed methods further comprise administering to the subject an antipsychotic medication if the subject is determined to be at risk of developing a psychotic disorder or is determined to be developing a psychotic disorder.

In some embodiments of the presently disclosed methods, the quantifying step comprises employing an plurality of antibodies, at least one of which binds to each of the gene products or other endogenous or exogenous substances with sufficient specificity to allow for quantification of the gene products or other endogenous or exogenous substances in the biological sample.

In some embodiments of the presently disclosed methods, the summary measure of expression for the subject is calculated by determining a z-score for each analyte assayed, wherein the z-score is based on the average and standard deviation of the unaffected comparison subjects, and summing the individual calculated z-scores.

In some embodiments of the presently disclosed methods, the subject is a female and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise (1) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E; or (2) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2; or (3) malondialdehyde- modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of calbindin 1, transforming growth factor beta, and cytokine (c-c motif) ligand 18; or (4)malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of calbindin 1, transforming growth factor beta, and cytokine (c-c motif) ligand 18. In some embodiments of the presently disclosed methods, the subject is a male and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise (1) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E; or (2) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, o2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2; or (3 )malondialdehy de- modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of interleukin 15 and chemokine (c-c motif) ligand 11; or (4) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin- IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2-macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of interleukin 15 and chemokine (c- c motif) ligand 11.

In some embodiments of the presently disclosed methods, the subject is not being treated with an antidepressant and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise (1) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin- IB, matrix metalloproteinase 7, and immunoglobulin E; or (2) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, 2-macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2; or (3) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of chromogranin A, endothelin 1, or both; or (4) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2-macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of chromogranin A, endothelin 1, or both.

In some embodiments of the presently disclosed methods, the subject is not being treated with an antipsychotic and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise (1) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-I B, matrix metalloproteinase 7, and immunoglobulin E; or (2) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-I B, matrix metalloproteinase 7, and immunoglobulin E, and one or more of n-ansthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2-macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2; or (3) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matiix metalloproteinase 7, and immunoglobulin E, and one or more of N-(alpha)-acetyltransferase 15, ferritin, and alpha 1-anti-chymotrypsin; or (4) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, ot2-macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of N-(alpha)-acetyltransferase 15, ferritin, and alpha 1-anti-chymotrypsin.

Thus, the presently disclosed subject matter provides methods for identifying a risk of developing a psychotic disorder in a subject and also for identifying whether or not a subject is developing a psychotic disorder.

BRIEF DESCRIPTION OF THE FIGURES

Figures 1A-1C provide a comparison of positive prediction for a diagnostic test in groups with low (1%; Figure 1A), medium (30%; Figure IB), and high (50%; Figure 1C) risk of disease. A "true positive" or "true negative" result means that the test was correct. In the three examples given in Figures 1A-1C, the test sensitivity (true positive/all actually positive) and specificity (true negative/all actually negative) were assumed to be 0.80. The positive predictive value (PPV; true positive/all who test positive) and negative predictive value (NPV; true negative/all who test negative) varied depending on the risk of disease in the tested population. Other test metrics, such as sensitivity, specificity, and accuracy, did not depend on the disease rate in the population of interest. In a cohort of 1000 patients with a low 1% risk of disease (see Figure 1A), the test identified 206 patients as "at-risk", but only 8 (4%) were true positives; the vast majority of persons with positive test results were in fact false positives. The same test becomes more clinically useful in a cohort of 1000 patients with moderate (30%) risk (see Figure IB). Here the test identified 380 persons as "at-risk", and 240 (63%) were true positives. In a cohort of 1000 patients at high (50%) risk (see Figure 1C), the test identified 500 as "at- risk", and 400 (80%) were true positives. The respective PVs were thus good, ranging from 80% to 99.7%. NPV would be expected to diminish, however, if the test were applied in patient populations with veiy high risk of disease (for example, patient populations with greater than 50%, 60%, 70%, or 75% risk).

Figure 2 is a flow diagram of an exemplary greedy algorithm for determining a plurality of analytes that could be employed for identifying a risk of developing a psychotic disorder in a subject.

Figure 3 is a graph of the results of 5x5x5-fold cross-validation of a greedy algorithm with random 80%-20% partitions of each group (subjects who met psychosis risk syndrome criteria and later developed a psychotic disorder (CHR-P), subjects who met psychosis risk syndrome criteria but later did not develop a psychotic disorder (CHR- NP), and unaffected comparison subjects who did not meet psychosis risk syndrome criteria (UC)). There were 125 permutations; thus, the maximum number of times that an analyte could be selected was 125. The more frequently selected analytes were, of course, considered to be more informative than the ones selected only 25 to 50 times.

Figures 4A and 4B are histograms of frequencies of values of the area under the resulting receiver operating curve (AUC) for pemiutations of random data. The types of the data were randomly permuted and were allocated into bins of 35, 40, and 32 samples (same as UC, CHR-NP, CHR-P). Then, classifiers from the pseudo data were built using the same greedy algorithm and five-fold cross validation process, retaining the sum of the five most frequently selected analytes in every trial. Both true data and 100 trials of pseudo data were thereby used in 101 classifiers with sums of the five most frequently chosen analytes. In terms of AUCs, the true data classifiers had higher performance than all (CHR-NP versus CHR-P) or all but one (UC versus CHR-P) classifier built with random data (and applied to the same random data). Figure 4A relates to AUCs for UC versus CHR-P of the classifiers built with just the five most selected analytes of 101 classifier constructions with 125 iterations of the greedy algorithm each, one with actual data and 100 with data with randomly permuted sample labels. In all trials, analytes could be chosen up to 125 times in five-by-five-by-five cross validations. Thus, the construction of the true classifier (applied to true data) outperformed the 100 pseudo classifiers (applied to respective pseudo data). Shown is a beta distribution fit and corresponding p- value. Figure 4B relates to the same exercise calculating 101 AUCs of the same 101 classifiers applied to CHR-NP versus CHR-P. Thus, in both sets of data the true classifier distinguished true data by choosing as first five analytes a combination with higher AUC values. The observations in this Figure, like those in Figure 5, indicated that: (1) the raw assay data must have had information distinguishing UC from CHR-P and CHR-NP from CHR-P; (2) the normalization method used did not obliterate the information; and (3) the classifier construction that was employed actually revealed the information.

Figures 5A and 5B are plots of the frequencies of first five analytes chosen with true labels and randomly permuted labels of the 107 subjects into subsets of 35, 40, and 32. The most frequently chosen analyte with randomly permuted labels was chosen less frequently than the most frequently chosen analyte with true labels, as were the second most frequently chosen, and so on. In Figure 5A, frequencies of the five most selected analytes of two classifier constructions with 125 iterations of the greedy algorithm each, one with actual data (black; plots 1, 3, 5, 7, and 9 from the left) and one with data with randomly permuted sample labels (gray; plots 2, 4, 6, 8, and 10 from the left). In both trials, analytes could be chosen up to 125 times in 20 five-by-five-by-five cross validations. The plots show the minimums, maximums, and middle quartiles of the frequencies over the 20 trials. Thus, the construction of the true classifier made more parsimonious use of the 141 possible analytes, that is, used fewer analytes more frequently in the first five chosen. This is a trait of classifiers that are not overfitted. Figure 5B is a comparison of the true classifier average total frequency with those of 100 pseudo classifiers. A pseudo classifier started with a random permutation of the 107 sample labels and type memberships and was otherwise constructed exactly like a true classifier. The pseudo classifiers were applied to the respective sets of pseudo data. A beta distribution fit the histogram generated by the total of the frequencies of the first five analytes used in the 100 pseudo classifiers. Thus, the true classifier distinguished true data by choosing as the first five analytes a combination with higher frequencies. The observations in this Figure, like those in Figure 4, indicated that: (1) the raw assay data must have had information distinguishing UC from CHR-P and CHR-NP from CHR-P; (2) the normalization method used did not obliterate the information; and (3) the classifier construction was employed actually revealed the information.

Figures 6A and 6B are receiver operating curves for the classifier constructed with all 107 samples from the 18 most frequently selected analytes in 125 five-by-five -by-five cross validations. Figure 6A pertains to the UC versus CHR-P classification and Figure 6B is the same classifier applied to the CHR-NP versus CHR-P classification. In both Figures 6A and 6B, the middle (darker) curve of each group of three curves represents the data generated and the first and third (lighter) curves represent the 95% confidence interval (CI).

Figure 7 is a schematic diagram of the Immune-Hypothalamus-Pituitary (IHP) interactions in blood analytes included in the predictive index. The asterisk indicates analytes that were included in the index.

Figure 8 is a graph illustrating reproducibility of assays for the 18 most frequently selected analytes. A technical replicate of the sample from subject 283 was assayed twice, each time in duplicate, to generate the normalized z-scores. Of the 141 analytes assayed for the duplicates, 111 reported values were not minima. Over the full 141 analytes the correlation was 0.84; for the 18 analytes identified in Figure 8, the correlation was 0.96 including three of the analytes for which both samples were at their minima; these 18 analytes are shown in Figure 8. Data were generated in duplicate for subject 283 to insure quality. The darker line in the Figure denotes the normalized values from one of the assays of subject 283 and the lighter line denotes the normalized values from the other assay of subject 283. The 18 analytes listed in Figure 8 are, from left to right, mucin-16, intei eukin-8, malondialdehyde-modified low density lipoprotein, matrix metalloproteinase-7, uromodulin, immunoglobulin E, growth hormone, chemokine (c-c motif) ligand 8, Factor 7, thyroid stimulating hormone, KIT ligand, Cortisol, interleukin- IB, resistin, apolipoprotein D, alpha-2-macroglobulin, chemokine (c-c motif) ligand 2, and transthyretin. It is noted that for the analytes mucin-16, interleukin-8, malondialdehyde-modified low density lipoprotein, matrix metalloproteinase-7, uromodulin, and immunoglobulin E, the data values were so nearly identical in each replicate that the traces are not depictable in the Figure as being distinct. Thus, the overlapping darker and lighter lines are depicted in the Figure as being a lighter line only.

DETAILED DESCRIPTION

The presently disclosed subject matter differs from previous studies in several respects. First, different patient populations were employed, including persons at elevated risk for psychosis, some of whom developed a psychotic disorder within two years, while others who did not. In addition, very different data analytic methods that are methodologically robust, avoid overfitting, and thus produce a reliable indicator of disease risk are disclosed herein. Importantly, the level of no single analyte provided meaningful sensitivity or specificity to increase psychosis risk prediction. Rather, a summary measure of a specific group of analytes that has high sensitivity and specificity for psychosis risk prediction is disclosed herein.

The selection of analytes used a greedy algorithm of polynomial (N²) complexity plus five-fold cross validation. To assure confidence in the selection, the entire method including the same algorithms was applied repeatedly to the same raw data after random permutation of subject labels. Note that this process is not to be confused with simply applying a single, finished classifier to permuted data, resulting trivially in distinguished performance for true data. The performance of the classifier built with true data was compared with that of such pseudo classifiers relative to pseudo data. As described below, these tests successfully demonstrated three achievements: the assay technology must have generated a signal; the normalization methods must not have obliterated the signal; and the analytic methods indeed revealed the signal. That is, the statistical outcome of comparing the true classifier built with true data with the pseudo classifiers could not reasonably be explained by chance and thus implied these three successes in our program.

The present subject matter will be now be described more fully hereinafter with reference to the accompanying Examples, in which representative embodiments of the presently disclosed subject matter are shown. The presently disclosed subject matter can, however, be embodied in different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the presently disclosed subject matter to those skilled in the art. All technical and scientific terms used herein, unless otherwise defined below, are intended to have the same meaning as commonly understood by one of ordinary skill in the art. References to techniques employed herein are intended to refer to the techniques as commonly understood in the art, including variations on those techniques or substitutions of equivalent techniques that would be apparent to one of skill in the art. While the following terms are believed to be well understood by one of ordinary skill in the art, the following definitions are set forth to facilitate explanation of the presently disclosed subject matter.

Following long-standing patent law convention, the terms "a", "an", and "the" mean "one or more" when used in this application, including the claims. Thus, the phrase "an analyte" refers to one or more analytes, unless the context clearly indicates otherwise.

The term "about", as used herein when referring to a measurable value such as an amount of weight, time, dose, etc., is meant to encompass variations of in some embodiments ±20%, in some embodiments ±10%, in some embodiments ±5%, in some embodiments ±1%, and in some embodiments ±0.1% from the specified amount, as such variations are appropriate to perform the disclosed methods.

As used herein, the term "and/or" when used in the context of a list of entities, refers to the entities being present singly or in combination. Thus, for example, the phrase "A, B, C, and/or D" includes A, B, C, and D individually, but also includes any and all combinations and subcombinations of A, B, C, and D.

The term "comprising", which is synonymous with "including" "containing", or "characterized by", is inclusive or open-ended and does not exclude additional, unrecited elements and/or method steps. "Comprising" is a term of art that means that the named elements and/or steps are present, but that other elements and/or steps can be added and still fall within the scope of the relevant subject matter.

As used herein, the phrase "consisting of excludes any element, step, or ingredient not specifically recited. For example, when the phrase "consists of appears in a clause of the body of a claim, rather than immediately following the preamble, it limits only the element set forth in that clause; other elements are not excluded from the claim as a whole.

As used herein, the phrase "consisting essentially of limits the scope of the related disclosure or claim to the specified materials and/or steps, plus those that do not materially affect the basic and novel characteristic(s) of the disclosed and/or claimed subject matter.

With respect to the terms "comprising", "consisting essentially of, and "consisting of, where one of these three terms is used herein, the presently disclosed and claimed subject matter can include the use of either of the other two terms. For example, the presently disclosed subject matter relates in some embodiments to comparing a summary measure of expression of a plurality of analytes in one subject to a standard, which in some embodiments can comprise a summary measure that is normalized with respect to healthy subjects of similar age, sex, and/or ancestry as the subject. It is understood that the presently disclosed subject matter thus also encompasses methods wherein the standard consists essentially of a summary measure that is normalized with respect to healthy subjects of similar age, sex, and/or ancestry as the subject, as well as methods that in some embodiments employs a standard that consists of a summary measure that is normalized with respect to healthy subjects of similar age, sex, and/or ancestry as the subject. Similarly, it is also understood that in some embodiments the methods of the presently disclosed subject matter comprise the steps that are disclosed herein and/or recited in any given claim, in some embodiments the methods of the presently disclosed subject matter consist essentially of the steps that are disclosed herein and/or recited in any given claim, and in some embodiments the methods of the presently disclosed subject matter consist of the steps that are disclosed herein and/or recited in any given claim.

Disclosed herein are methods that can be employed for identifying a risk of developing a psychotic disorder in a subject and/or for whether or not a subject is developing a psychotic disorder. The presently disclosed methods comprise in some embodiments (a) isolating and/or providing a biological sample comprising serum, plasma, urine, and/or saliva from a subject who meets clinical criteria for elevated risk of psychosis; (b) quantifying a level of expression in the biological sample of a plurality of gene products or other endogenous or exogenous substances, or a nuxture thereof, normally found in the biological sample, wherein the gene products or other endogenous or exogenous substances, or the mixture thereof, are selected from the group consisting of malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleuldn-IB, matrix metalloproteinase 7, and immunoglobulin E; (c) combining the quantified levels of expression from step (b) to create a summary measure of expression; and (d) comparing the summary measure of expression for the subject to one or more standards, wherein the comprising step identifies a risk of developing a psychotic disorder in the subject.

As used herein, the phrase "clinical criteria for elevated risk of psychosis" refers to the 19 psychological criteria outlined in the "Scale of Prodromal Symptoms" (SOPS; see Miller et at, 2003). These 19 psychological criteria include Positive Symptoms (Unusual Thought Content/Delusional Ideas, Suspiciousness/Persecutory Ideas, Grandiosity, Perceptual Abnormalities/Hallucinations, and Disorganized Communication), Negative Symptoms (Social Anhedonia, Avolition, Expression of Emotion, Experience of Emotions and Self, Ideational Richness, and Occupational Functioning), Disorganization Symptoms (Odd Behavior and Appearance, Bizarre Thinking, Trouble With Focus and Attention, and Personal Hygiene), and General Symptoms (Sleep Disturbance, Dysphoric Mood, Motor Disturbances, and Impaired Tolerance to Normal Stress). Each symptoms is given a rating of severity (0, 1, 2, 3, 4, 5, 6) for criteria such as "paranoia" and "inability to plan."

As used herein, the phrase "a plurality of gene products or other endogenous or exogenous substances, or a mixture thereof refers generally to a plurality of analytes the presence of which can be assayed in a biological sample isolated from a subject such as but not limited to blood. As used herein, the term "analyte" refers to any measurable substance that is present in a biological sample from a subject, an analysis of the concentration and/or expression of which in the biological sample can be employed in the methods disclosed herein to identifying subjects at risk for developing a psychotic disorder and/or whether or not a subject is developing a psychotic disorder. An analyte can be a protein, peptide, lipid, and/or carbohydrate, and/or any modified variant thereof, alone or in any combination, which is present in a biological sample isolated from a subject including, but not limited to a human subject.

As used herein, the phrase "endogenous or exogenous substances" refer to anything that might be present in and therefore assayable with respect to expression level, concentration, etc. in a biological sample that can be isolated from a subject. "Endogenous" substances are those that are produced by or otherwise originate from within the subject itself. By way of example and not limitation, an endogenous substance can be a gene product encoded by the subject's genome, or a modified variant thereof. In some embodiments, the modified variant is modified by a biological process that occurs within the subject, such as but not limited to malondialdehyde-modified low density lipoprotein, which is a lipid peroxidation marker of oxidative stress associated with chronic stress and inflammation. Each of the gene products and other analytes listed in Tables 1-5 are considered "endogenous substances" as that term is being employed herein.

With respect to "exogenous" substances, these include any substance that is found within a biological sample but that has been introduced into the subject, either intentionally (including but not limited to ingestion, administration, etc.) or unintentionally (e.g., by infection of a pathogen). Food and therapeutic agents are considered "exogenous agents", as are viruses, bacteria, etc.

As used herein, the phrase "biological sample" refers to any material that can be isolated from a subject and that is expected to comprise an analyte. Exemplary, non- limiting biological samples include a body fluid (for example blood, cerebral spinal fluid, saliva, urine, etc., or any fraction thereof such as, but not limited to plasma), a cell (for example white blood cells, red blood cells, cultured human cells, etc.), and a tissue (for example skin, fat, olfactory epithelium, bone marrow, etc.). Within an individual subject, the quantity of an analyte can vary, depending on the biological sample from or in which it is measured, the time of day that it is measured, any recent use of a drug and/or ingestion of food, physical exercise, and/or any other factors that might impact a level of an analyte that might be present in an individual.

Among individuals, factors such as age, gender, and/or ancestry can also impact the level of an analyte. Optimally, the levels of analytes employed as predictors and/or indicators of developing a psychotic disorder should be mainly impacted by disease risk and minimally impacted by other factors not related to disease risk. In some embodiments, a biological sample comprises human blood, and in some embodiments the blood specimen is collected mid-day. In some embodiments, the blood is collected in tubes that contain compounds to protect the stability of analytes present within the blood, particularly those to be assayed in the presently disclosed methods. In some embodiments, blood isolated from a subject is processed to obtain the biological sample to be assayed, which is then tested for levels of expression of desired analytes.

In some embodiments of the presently disclosed methods, the subject is not being treated with an antidepressant when a biological sample is isolated from the subject. The phrase "not being treated with an antidepressant" refers to the subject not receiving any antidepressant therapeutic agent for a time period that is sufficiently long that any therapeutic effect and/or any effect on the expression of any analyte employed in the practice of the presently disclosed subject matter would be expected to have been eliminated in the subject. Exemplary time periods for clearance of antidepressants can be in some embodiments 24 hours, in some embodiments 48, hours, in some embodiments 72 hours, in some embodiments one week, in some embodiments two weeks, in some embodiments one month, in some embodiments two months, in some embodiments three months, in some embodiments six months, etc. For subjects that are not being treated with an antidepressant, the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample can in some embodiments further comprise chromogranin A, endothelin 1 , or both.

Similarly, in some embodiments of the presently disclosed methods, the subject is not being treated with an antipsychotic. The phrase "not being treated with an antipsychotic" refers to the subject not receiving any antipsychotic therapeutic agent for a time period that is sufficiently long that any therapeutic effect and/or any effect on the expression of any analyte employed in the practice of the presently disclosed subject matter would be expected to have been eliminated in the subject. Exemplary time periods for clearance of antipsychotics can also bein some embodiments 24 hours, in some embodiments 48, hours, in some embodiments 72 hours, in some embodiments one week, in some embodiments two weeks, in some embodiments one month, in some embodiments two months, in some embodiments three months, in some embodiments six months, etc. For subjects that are not being treated with an antipsychotic, the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample can in some embodiments further comprise one or more of N-(alpha)-acetyltransferase 15, ferritin, and alpha 1-anti-chymotrypsin.

As used herein, the phrase "level of expression" relates to an amount of an analyte that is present in a biological sample at a time at which it is assayed. Levels of expression can be referred to in whatever measure is desirable. Exemplary measures for levels of expression include absolute measures such as concentration {e.g. , nanograms/microliter or any other such mass per unit volume or mass per unit mass determination), relative measures, or any other measure that provides a reasonably repeatable articulation of an amount of an analyte in a biological sample. Those skilled in the art recognize the variety of methods available to measure the expression level of an analyte from a bodily fluid, tissue extract, cell extract, etc. For gene products, a level of expression can relate to an abundance of a transcription or translation product of that gene, which can be determined using standard molecular biological techniques including, but not hmited to quantitative reverse transcription polymerase chain reaction (qRT-PCR), enzyme-linked immunosorbent assay (ELISA), mass spectrometry, planar microarrays (protein chips), and bead-based microarrays (suspension arrays; see Jun et ah, 2012). These methods and others can be used alone or in combination with any suitable method of measuring the levels of the specified analytes. The choice of method for measuring the expression levels of an analyte or combination of analytes can be made by evaluating any relevant property of the test, including but not limited to the inherent variability of the test result, the test dynamic range, the amount of sample required, the complexity of the test protocol, and/or any other practical consideration that might be relevant to a particular quantifying methodology.

It is important to appreciate that in some embodiments, the biomarker tests reflected in the steps of the presently disclosed methods can be most useful when performed on persons that have aheady been determined to be at elevated risk for psychosis, typically based on clinical criteria. This is related to the fact that the proportion of patients who test positive on a given biomarker test who actually get the disease (positive predictive value; PPV) can vary dramatically depending on the actual rate of disease in the tested population. If the rate of disease is low, for example if only 1% of tested persons are actually at risk for the disease, than even a test with high sensitivity and specificity (-0.8) can have a significant fraction of false positives as true positives (see Figure 1). Thus, to minimize the risk of misdiagnosis of disease risk and the risk of embarking on an inappropriate treatment program, the usefulness of any biomarker test with less than perfect specificity is greatest in a relatively high risk population, for example where at least 10% of persons will ultimately develop the disorder.

There are various clinical criteria that are now established that indicate a person is at high risk for the development of a psychotic disorder. In some embodiments, at least a 10-fold elevation in risk as compared to the population at least is considered a "high risk". Examples include persons meeting the Criteria for Prodromal States (COPS), ultra high- risk criteria, basic symptom criteria, schizotypal personality disorder criteria or having a relative with a psychotic disorder (Golembo-Smith et ah, 2012; Schultze-Lutter et αί, 2013; see also Miller et al, 2003; Lencz et al, 2004; Fusar-Poli et al, 2012). A recent meta-analysis of studies that in total included about 2500 persons meeting clinical high risk for psychosis criteria (CHR) concluded that transition to psychosis rates at 1 year were -22%, at 2 years were -29%, and at 3 years were -32% (Fusar-Poli et al, 2012). In addition, genetic factors can identify persons at elevated risk for psychosis. For example, different lines of research indicate that persons with a specific genomic copy number variation, 22ql l syndrome, have a 30-fold increase in risk of developing a psychotic disorder compared to the general population risk (Green et al, 2009). As research advances it is conceivable that other combinations of various common and/or rare polymorphisms might be identified that predict a clinical high risk state.

Other clinical or biological assessments can further indicate even greater elevations in psychosis risk in persons meeting clinical or other criteria for elevated psychosis risk. Examples include, but are not limited to, neurocognitive impairments (Seidman et al, 2010), social cognitive impairments (Healey et al, 2013), social and vocational functional impairments (Cornblatt et al, 2012), salivary Cortisol levels (Walker et al, 2013), change in gray or white matter volume (Witthaus et al, 2008; Chan et al, 2009; Takahashi et al, 2009), electroencephalogram (Shin et al, 2009; Belger et al, 2012; Nagai et al, 2013; Perez et al, 2013) and so on. Therefore, in some embodiments a blood test can be combined with other measures of psychosis risk vulnerability to increase psychosis prediction.

In practice, a person using a biomarker test should ideally be aware of the actual disease risk in the person tested in order to correctly interpret test results related to psychosis risk. For example, if the test is used in a general population, different cut-off points with high specificity (> 0.99) at the sacrifice of sensitivity can be desirable. The presently disclosed methods are thus in some embodiments used in persons who meet criteria for at minimum a -10-fold increase in psychosis risk compared to the general population risk of -1% as assessed using criteria other than the presently disclosed methods. For example, if the risk of psychosis in the tested population is 30% within two years, then in such a population a biomarker test with good sensitivity and specificity (~ 0.80) can achieve PPV of about 0.63; equivalently, about two-thirds of persons identified by such a test as at risk would truly be on a trajectory to develop psychosis. Of perhaps equal significance is that such a diagnostic test would have a negative predictive ability of -0.90, greatly improving a clinician's confidence in predicting who is likely not to develop psychosis. For this reason it is conceivable that blood assays and/or other biomarkers with good sensitivity and specificity can have clinical utility in enhancing prediction of psychotic affliction, identifying among patients presenting with prodromal signs and symptoms those persons where concern for psychosis is greatest as well as those for whom concern for psychosis is actually quite low.

One a subject is identified who is at risk for developing a psychotic disorder and/or is presently developing a psychotic disorder, a treatment methodology can be implemented wherein the subject undergoes some therapeutic treatment to address the psychotic disorder. Treatment strategies for addressing psychotic disorders are known to those of skill and include, but are not limited to administration of antipsychotic medication (such as, but not limited to chlorpromazine, flupenthixol, fluphenazine, haloperidol, loxapine, perphenazine, pimozide, thioridazine, thiothixene, trifluoperazine and zuclopenthixol); and psychosocial intervention including, but not limited to patient case management, supportive psychotherapy, group therapy, individual Cognitive Behavior therapy (CBT), and/or vocational counselling. In some embodiments, a subject who is identified to at risk for developing a psychotic disorder and/or is presently developing a psychotic disorder using the methods disclosed herein is placed on antipsychotic medications and/or given one or more types of supportive psychosocial interventions.

By way of example and not limitation, a subject can present with clinical high risk symptoms. A blood specimen can be collected and assayed for the presently disclosed markers. Initially the five "core" analytes, malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, interleukin-lb, matrix metalloproteinase 7, and immunoglobulin E, are assayed. Optionally, one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (C-C motif) ligand 8, Factor 7, Cortisol, resistin, a2-macroglobulin, mucin- 16, chemokine (C-C motif) ligand 2 are also assayed. Other characteristics of the subject can also be employed to add further analytes to the assay, such as whether the subject is female (one or more of calbindin 1, transforming growth factor beta, and cytokine (c-c motif) ligand 18 can be added), male (one or more of interleukin 15 and chemokine (c-c motif) ligand 11 can be added), the subject is not being treated with an antidepressant (chromogranin A and/or endothelin 1 can be added), and/or the subject is not being treated with an antipsychotic (one or more of N-(alpha)-acetyltransferase 15, ferritin, and alpha 1-anti-chymotrypsin can be added).

The result of the assays will yield a new vector of values (in some embodiments, a normalized component-by-component as disclosed herein). The classifier function disclosed herein is then applied to the new vector. A clinician can then choose a combination of specificity and sensitivity (i.e., a compromise or trade-off is chosen based on relevant considerations for that subject) on the ROC curve, hence a threshold (break point) for the classifier function. For example, the clinician might regard false negatives as more costly per subject than false positives, hence a choice is made to reduce false negatives. From this, a decision with respect to treatment is chosen.

Alternatively, the new vector can compared to some or all of the historical vectors that can be generated are from persons who did progress to schizophrenia. The comparison could be among mean calculations of Pearson correlations or Spearman correlations and/or other known method of comparing n-dimensional vectors. The new vector can also be compared to all the historical vectors that are from persons who did not progress to schizophrenia. In such an embodiment, the presently disclosed subject matter can compare all of the comparisons to determine whether the new vector is more similar to the PROGRESSED patients or the NOT PROGRESSED patients.

For example, the new vector can have Pearson correlations with 1000 PROGRESSED vectors with an average of 0.1 and a standard deviation of 0.2. The new vector can also have Pearson correlations with 1000 NOT PROGRESSED vectors with an average of 0.5 and a standard deviation of 0.2. This would provide strong evidence that the new vector represents a subject who is unlikely to progress to schizophrenia or another psychotic disorder. A clinician could then decide not to embark at present on a regime of treatments that would themselves inevitably carry risks of adverse side-effects.

It is noted that comparisons of new vectors with two or more sets of historical vectors can be accomplished using techniques that are well known to those skilled in the art.

The presently disclosed subject matter also provides assay kits comprising reagents, in some embodiments antibodies, for detecting the presence of and quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances, or mixtures thereof, normally found in the biological samples, wherein the gene products or other endogenous or exogenous substances, or the mixture thereof, are selected from the group consisting of malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E. the reagents can be, in some embodiments, antibodies that binds to malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E. In some embodiments, the kits further comprise reagents that bind to one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2-macroglobulin, mucin- 16, chemokine (c-c motif) ligand 2, of beta 2 transferrin, prostaglandin D synthase (beta trace protein), adrenocorticotropin releasing hormone, insulin-like growth factor, chromogranin A, endothelin 1, N-(alpha)- acetyltransferase 15, ferritin, alpha 1-anti-chymotrypsin, calbindin 1, transforming growth factor beta, cytokine (c-c motif) ligand 18, interleukin 15, and chemokine (c-c motif) ligand 11. In some embodiments, the reagents for detecting the presence of and quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances hsted herein above are detectably labeled. In some embodiments, the kits further comprise compounds that permit detection of the detectably labeled reagents. In some embodiments, one or more of the reagents provided in the kit is labeled with a different detectable label, allowing for simultaneous detection and quantification of multiple gene products or other endogenous or exogenous substances in biological samples.

In some embodiments, the reagents for detecting the presence of and quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances are antibodies, optionally monoclonal antibodies. In some embodiments, the antibodies are affixed to a solid support.

In some embodiments, the reagents for detecting the presence of and quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances are oligonucleotide probes that are specific for one or more of the gene products or other endogenous or exogenous substances. In some embodiments, the nucleotide sequences of the oligonucleotide probes are designed to bind to cDNAs derived from the gene products but not bind to genomic DNA (i.e., have sequences that flank introns that are present in the subject's genome such that the oligonucleotides include sequences from different exons of the gene products. EXAMPLE

The following Example provides further illustrative embodiments. In light of the present disclosure and the general level of skill in the art, those of skill will appreciate that the following Example is intended to be exemplary only and that numerous changes, modifications, and alterations can be employed without departing from the scope of the presently disclosed subject matter.

Material and Methods for the EXAMPLE

Subjects. The aims and methods of the North American Prodrome Longitudinal Study (NAPLS2) were described in detail in Addington et al, 2012. To summarize, NAPLS2 was an eight-site observational study of the predictors and mechanisms of conversion to psychosis in persons meeting the Criteria of Prodromal State (COPS; Miller et al, 2002). The study took place at Emory University, Harvard University, University of Calgary, University of California Los Angeles, University of California San Diego, University of North Carolina Chapel Hill, Yale University, and Zucker Hillside Hospital, and it included a total of 720 clinical high risk (CHR) subjects and 240 demographically similar, unaffected comparison (UC) subjects. All subjects were between the ages of 12 and 35 years old, and were followed for up to two years. The study was approved by the Institutional Review Board at each site, and each subject provided written informed consent or assent, and parent or guardian also consented for all minor subjects.

107 of the subjects from NALPS2 were employed in the development of the presently disclosed subject matter. Thirty-two of these subjects met psychosis risk syndrome criteria and later developed a psychotic disorder (referred to herein as "CHR- P"), 40 met psychosis risk syndrome criteria but did not develop psychosis during 24 month follow-up (referred to herein as "CHR-NP"), and 35 unaffected comparison subjects did not meet psychosis risk syndrome criteria (referred to herein as "UC"). The CHR-P and CHR-NP subjects included all those available with plasma samples by February 2012. The UC subjects were chosen to be demographically similar to the CHR subjects. The study exclusion criteria for CHR and UC participants included meeting DSM-IV criteria for an Axis I psychotic disorder; also excluded were candidates for UC membership having a first-degree relative with a currerit or past psychotic disorder. General exclusion criteria were having a substance dependence (past 6 months), neurological disorder, or Full Scale IQ < 70. Assessments. Participants were screened using the Structured Interview for Prodromal Syndromes (SIPS; Miller et al, 2002) for the presence of one or more prodromal syndromes: attenuated, subthreshold psychotic symptoms, brief intermittent psychotic symptoms, substantial functional decline combined with first degree relative with a psychotic disorder, or schizotypal personality disorder in individuals less than 18 years old. The SIPS was administered at initial assessment and follow-ups.

The SIPS contained the Schedule of Prodromal Symptoms (SOPS), which rated the severity of relevant symptoms with the following scale: 0 = absent; 1 = questionably present; 2 = mild; 3 = moderate; 4 = moderately severe; 5 = severe but not psychotic; and 6 = severe and psychotic. The SOPS was composed of four symptom domains that were classified as positive (e.g., unusual thought content, suspiciousness, grandiose ideation, perceptual abnormalities, disorganized communication); negative (e.g., social anhedonia, avolition, expression of emotion, experience of emotions and self, ideational richness, occupational functioning); disorganized (e.g., odd behavior or appearance, bizarre thinking, trouble with focus and attention, impairment in personal hygiene); and general (sleep disturbance, dysphoric mood, motor disturbances, impaired tolerance to normal stress).

There were three criteria for clinical high risk: attenuated positive symptom; genetic risk and deterioration; and brief intermittent psychotic symptom. To meet attenuated positive symptom criteria, subjects had a rating of "3", "4", or "5" on at least one of the positive symptom items and at least one symptom that began or worsened in the past year and occurred at least once per week in the past month. To meet genetic risk and deterioration criteria, subjects had functional decline defined as a 30 point decrease in rating by the Global Assessment of Functioning (GAF; Coffey et al, 1996) scale in the past year and schizotypal personality disorder or a first-degree relative with a psychotic disorder. To meet brief intermittent psychotic symptom criteria subjects had a rating of "6" on at least one of the positive symptoms items, lasting for several minutes a day at least once per month over the past three months. Psychosis criteria were met if the subject had a score of "6" on any of the SOPS positive items occurring for at least one hour per day at least four days in a month, with the symptoms seriously disorganizing or dangerous.

The Structured Clinical Interview for Axis I DSM-TV Disorders (SCID-I/NP; First et al, 2002) was administered during the initial evaluation and during subsequent annual follow-up assessments. The SCID-I/P was utilized to maintain consistency in the diagnostic procedure across participants and over time as they entered young adulthood through the longitudinal course of the study.

Clinical evaluations were conducted by interviewers who were trained to meet reliability standards for the project. Interviewers were clinical psychologists, psychiatrists, and other mental health professionals. Training of interviewers was conducted over a 2-month period; inter-rater reliability for symptoms ratings exceeded the minimum criterion of 0.80 (Pearson correlation); for diagnostic status, average Cohen's kappa was 0.85. Socioeconomic status was estimated by maximum years of education of mother or father. Participants were assessed clinically every six months over a two-year follow-up period. Biomarker assays were conducted at one year and two years after baseline, or at conversion to psychosis. Assays included brain imaging, electrophysiology, saliva collection, blood collection, and neurocognitive testing.

Plasma Collection. Blood samples were collected in Becton Dickenson PI 00 blood collection tubes that contain EDTA as anticoagulant, proprietary protein stabilizers, and a mechanical separator. Most samples were processed within 2 hours, and the plasma stored at -80°C until analysis.

Plasma Assay. Plasma samples were sent on dry ice to Myriad Rules Based Medicine, a biomarker testing laboratory that has maintained Clinical Laboratory Improvement Amendments (CLIA)-accreditation by COLA (Columbia, Maryland, United States of America) since 2006. Samples were analyzed with the Human DISCOVERYMAP® assay (Myriad RBM, Austin, Texas, United States of America), a LUMINEX® bead-based multiplex immunoassay (Luminex Corporation, Austin, Texas, United States of America) that included 185 analytes involved in hormonal responses, inflammation, growth, oxidative stress, and metabolism. The value for an individual analyte was based on a standard curve, and samples were run in duplicate. The least detectable dose (LDD) was the concentration interpolated by the average plus 3 standard deviations of 20 readings of diluent blanks. The lower limit of quantification (LLOQ) was the lowest concentration of an analyte in a sample that could be reliably detected, as defined by the coefficient of variation of replicate standard samples < 30% (90% of analytes had a coefficient of variation of standard samples <15%). Technicians ran protein assays without knowledge of clinical status of the subjects and used standard protocols. Analyte values that were not quantifiable were converted to the LLOQ. Exclusion of Analytes. The original data set contained 185 analytes. Twenty-three analytes that were not detected in at least 20% of the subjects were excluded. Most of the included analytes (80%) were detected in at least 90% of the subjects.

Persons exhibiting CHR symptoms are often prescribed various medications by their community health care provider (Cadenhead et al, 2010), implying that medication use can confound differences in analyte expression comparisons between groups. At the times of blood draws, 48% of the CHR-NP subjects (19/40) and 38% of the CHR-P subjects (12/32) were on at least one prescription medication. The type of medication was as follows: 25% of CHR-NP subjects and 13% of CHR-P subjects were on an antipsychotic; 30% of CHR-NP subjects and 25% of CHR-P were on an antidepressant; 8% of CHR-NP subjects and 6% of CHR-P subjects were on a stimulant; 5% of CHR-NP subjects and 3% of CHR-P subjects were on a mood stabilizer; and 5% of CHR-NP subjects and 6% of CHR-P subjects were on a benzodiazepine. No subjects were taking a non-steroidal anti-inflammatory drug (NSAID) or an antibiotic at the time of the blood draw. Among the UC subjects, one was prescribed an antidepressant after enrollment but before the blood draw.

Given the small sample size, analytes with a relation to prescription of antipsychotics or antidepressants were excluded by using a non-adjusted Wilcoxon nonparametric test p- value < 0.05. Consequently, 21 more analytes were excluded (Table 1) for primary analyses. Secondary analyses included the full list of 168 analytes.

Table 1

Analytes with Possible Relationship to Antipsychotic or Antidepressant Prescription in

Clinical High Risk (CHR) Subjects*

Granulocyte-Macrophage Colony P04141 0.023 0.062 Stimulating Factor

Interferon Gamma P01579 0.024 0.514

Fatty Acid-Binding Protein, Heart P05413 0.026 0.610

Clusterin PI 0909 0.028 0.336

Complement Factor H Related Q03591 0.029 0.081 Protein

Cancer Antigen 19-9 Q9BXJ9 0.033 0.249

Matrix Metalloproteinase 2 P08253 0.040 0.027

Neuronal Cell Adhesion Molecule Q92823 0.045 0.358

Advanced glycosylation end product- Q15109 0.049 0.297 specific receptor

Proinsulin, Total P01308 0.622 0.011

Proinsulin, Intact P01308 0.288 0.013

Angiopoietin-2 015123 0.097 0.026

Insulin-Like Growth Factor Binding P18065 0.179 0.029 Protein 2

Chromogranin A PI 0645 0.955 0.036

Apolipoprotein B P04114 0.080 0.040

Interleukin 6 Receptor P08887 0.146 0.040

Proinsulin, Total P01308 0.131 0.040

Alpha-Fetoprotein P02771 0.340 0.043

Interleukin 10 P22301 0.086 0.048

* excluded from main analysis.

# Kruskal- Wallace p- alue.

Regarding frequencies of LLOQ in the 141 analytes, 70 had only one sample with the minimum value and another 32 had five or fewer instances of minimum values. The remaimng 39 analytes had six to 83 samples with minimum values. Four of the 39 are described herein to be especially informative: namely, malondialdehyde-modified low- density lipoprotein (64 minima), interleukin- 1 beta (63 minima), immunoglobulin E (11 minima), and interleukin-8 (13 minima).

Normalization. Raw data were used for permutation analyses. The normal plasma concentrations varied analyte-to-analyte up to 1,000,000-fold. For other analyses, so that results could be viewed on the same scale, the results for each analyte were standardized (z-score) to the average and standard deviation (sd) values of the unaffected comparison subjects. Thus, normalized average and standard deviation values for UC subjects were 0.00 and 1.00. With larger numbers of UC subjects, it would be feasible and could be desirable to take into account age, sex, ancestry, and/or drug/medication use, to name a few examples, when doing normalization.

Reproducibility. Regarding reproducibility, one sample was submitted twice (hence duplicate assays of duplicate samples) without the knowledge of the assaying laboratory. Of the 141 analyte assays for the duplicates, 111 reported values were not minima. Over the full 141 analytes their correlation was 0.84; for the 18 analytes (see Figure 8) described herein, the correlation was 0.96 including three of the analytes for which both samples were at their minima.

Data Analyses. Statistical analyses were performed using the Microsoft EXCEL® brand spreadsheet program (Microsoft, Redmond, Washington, United States of America), U SCRAMBLER® (CAMO Software AS, Oslo, Norway), or SAS® (SAS, Cary, North Carolina, United States of America) statistical software. Student t-tests or chi-square tests were used to compare the CHR and UC groups on demographic and clinical characteristics. For group comparisons of analyte levels, EASYFIT® software (MathWave Technologies, Dnepropetrovsk, Ukraine) was first used to test for rejection of Gaussian (normal) distribution by Kolmogorov-Smirnov (K-S) and Anderson-Darling (A- D) tests. The add-in SIGNIFICANCE ANALYSIS OF MICROARRAYS® (SAM; Tusher et al, 2001) for the Microsoft EXCEL® brand spreadsheet program was used to compare expression between groups with permutation, using the Wilcoxon statistic and 1000 permutations.

It was expected that a classifier would deliver good area under the resulting receiver operating curve (AUC) values (see discussion herein below) for various random 80% training subsets of the three types and also for the complementary 20% testing subsets. It was also expected that randomly reassigning the same numbers of samples to the three types and applying the very same classifier construction would result in pseudo- classification that would almost always be easily distinguished from true classification. The general principles and pitfalls of classifier construction were also noted (see Hand, 2006; Moons et al, 2012a; Moons et al, 2012b). Several types of standard classifiers were: linear methods (e.g., linear discriminant analysis, principal component analysis (PCA), and types of regression analyses); and complex algorithms with hidden terms and functions (support vector machines (SVMs) with four types of basis functions and many types of clustering algorithms). Results were good or excellent on training subsets but disappointing on testing subsets. For example, regarding principal component analysis, the explained total variances for calibration and validation were only 9.7% and 5.2%. These values were little better than that which would result from random shuffling of entries in the entire 107-by-141 matrix of z-scores, namely, about 4.2% and -0%. Thus PCA failed to discern any significant axis for the data. As another example, SVMs performed perfectly or nearly so on 80% training subsets of the three types, but classification of the other 20% subsets was essentially random.

This led to yet another type of classifier construction. Greedy algorithms are capable of selecting collectively informative markers from large candidate sets (Liu et al. , 2005). They linearly build marker selections and avoid brute force examination of all possible subsets of markers. A program that first selected the very best single analyte for distinguishing the three types was developed. Then a second analyte was added that best improved performance, if possible. Additional analytes were selected and added until no further selection of any analyte improved performance. The selections were made over numerous subsets of the subjects, and selected sets of analytes were intersected to find analytes that consistently contributed to performance (see Figure 2).

The chosen set of analytes created an "index", also referred to herein as a "summary measure of expression", defined as the sum of the z-scores of all selected analytes. In some embodiments, the index or summary measure takes into account the five "core" analytes (i.e., malondialdehy de-modified low density lipoprotein, thyroid stimulating hormone, interleukin-lB, matrix metalloproteinase 7, and immunoglobulin E), and in some embodiments the index or summary measure takes into account the 18 analytes listed in Figure 8. When analytes beyond the five core analytes are included in the analysis (for example, if the subject is not being treated with an antidepressant, one or both of the analytes chromogranin A and/or endothelin 1 can be added; if the subject is not being treated with an antipsychotic, one or more of the analytes N-(alpha)- acetyltransferase 15, ferritin, and alpha 1-anti-chymotrypsin can be added; if the subject is a female, one or more of the analytes calbindin 1 , transforming growth factor beta, and cytokine (c-c motif) ligand 18 can be added; and if the subject is a male, one or both of the analytes interleukin 15 and/or chemokine (c-c motif) ligand 11 can be added). In effect, this was a weighted sum with all weights being 1 ; this was possible because most of the analytes that moderately distinguished the three types tended to have higher values in CHR-P than CHR-NP and higher values in CHR-NP than UC. The resulting classifying function was therefore veiy simple. Performance was defined to be minimization of half the sum of the squares of the Student t-test p-values for CHR-P versus CHR-NP and CHR-P versus UC. Performance thus entailed distinguishing subjects who developed psychosis from both high risk subjects who did not progress to schizophrenia and unaffected control subjects.

In more detail, classifier analyses on the results of the greedy algorithm were executed using five-by-five-by-five-fold cross validation (Kohavi, 1995; Tropsha, 2010) with repeated random sub-sampling; it was implemented in the Microsoft EXCEL® brand spreadsheet program with macros (nested loops designed to implement the greedy algorithm depicted in Figure 2 and add-ins (Ablebits, a a project of Add-in Express Ltd., Homel, Belarus; CAMO Software Inc., Woodbridge, New Jersey, United States of America; MEDCALC® Software bvba, Ostend, Belgium; and MathWave Technologies, Dnepropetrovsk, Ukraine). The groups of 32 CHR-P, 40 CHR-NP, and 35 UC subjects were randomly assigned to 1 of 5 equal subgroups. In each of 125 combinations, 4 of the CHR-P, 4 of the CHR-NP, and 4 of the UC subgroups were used to train a classifier that was tested on the complementary subgroups. The performance of the 125 preliminary classifiers for the 125 tests was noted. The entire process was repeated 20 times: that is, 20 initial selections of the random 20% subgroups, for a total of 2500 executions of the greedy algorithm. A review of all the preliminary classifiers led to a final, integrated classifier because in the 125 tests several analytes were repeatedly selected while other analytes were never selected.

The area under the resulting receiver operating curve (AUC) was used to assess the capacity of the index to distinguish CHR-P from the other two groups using the exact (no smoothing) Hand-Till method (see Hand, 2010), with calculations of standard error following the methods of Handley-McNeil (Handley & McNeil, 1982). The AUC was, after the Student t-test /?-value, a second, reasonable performance metric, and two were used to avoid the risk of some sort of idiosyncratic effect of choosing one. A receiver operating curve is a plot of sensitivity (i.e., test correctly predicts positive/all true positives) versus the false positive rate (i.e., 1 — specificity; test correctly predicts negative/all true negatives). Various threshold settings yield the points along the curve. An AUC of 0.5 indicates that the classification is equal to chance, and an AUC of 1 indicates perfect classification.

Participant Characteristics. The clinical high risk (CHR) subjects that did not progress to psychosis within a two year follow-up period (CHR-NP), clinical high risk subjects who did progress to psychosis (CHR-P), and unaffected comparison (UC) subjects did not differ with respect to age, sex, ancestry, symptom severity, substance use, or time of blood draw, based on t-test/Chi-square test (see Table 2). All of the CHR subjects met attenuated positive symptom diagnostic criteria. The psychosis diagnoses of the CHR-P included 13 with psychosis, not otherwise specified, 14 with schizophrenia, two with major depression with psychotic features, and one each with schizoaffective, delusional, or bipolar disorder.

Table 2

Demographic and Clinical Characteristics of Study Subjects

SOPS Scores, average (sd)

Total¹ 5.06 (5.11) 36.63 (13.03) 43.81 (14.11)

Positive¹ 1.46 (1.84) 12.28 (4.74) 14.22 (3.92)

Negative¹ 1.23 (1.72) 11.38 (6.32) 13.58 (6.23)

Disorganized¹ .91 (1.17) 5.05 (2.79) 6.45 (3.88)

General¹ 1.46 (1.80) 7.93 (4.46) 10.52 (4.44)

Calgary Depression Scale for .89 (1.71) 5.00 (4.72) 6.88 (4.88)

Schizophrenia Scores', average

(sd)

Zung Self-Rated anxiety Scale², 28.79 (4.15) 45.80 (13.35) 48.39 (12.55) average (sd)

Time blood draw, average (sd) 12.20 (1.85) 12.65 (2.0) 11.97 (1.79)

Prescription Medication

Antipsychotic³ 0% 25% 13%

Antidepressant⁴ 3% 30% 25%

Stimulant 0% 8% 6%

Mood stabilizer 0% 5% 3%

Benzodiazepine 0% 5% 6%

NSAID 0% 0% 0%

Antibiotic 0% 0% 0%

Substance Use

Tobacco use⁵ 9% 33% 44%

Alcohol use 46% 48% 38%

Marijuana use⁶ 9% 28% 31%

'CHR-P vs. UC t-Test : ^ T^vaTue < 0.0001; CHR-NP vs. UC t-test = -7.56,_p-value <0.0001 ²CHR-P vs. UC t-test = -8.31 p-va ie < 0.0001; CHR-NP vs. UC t-test = -7.56, Rvalue O.0001 ³CHR-P vs. UC FET -value = 0.047; CHR-NP vs. UC FET /j-value == 0.001

4CHR-P vs. UC FET/j-value = 0.011; CHR-NP vs. UC FET Rvalue = 0.002

5CHR-P vs. UC FET Rvalue = 0.002; CHR-NP vs. UC FET -value = 0.022

6CHR-P vs. UC FET/>-value = 0.029; CHR-NP vs. UC FET /j-value = 0.042

Plasma Analytes and Development of Psychosis. Controlling for multiple comparisons, there were no analytes that were determined by Wilcoxon/Student's t-test to be differentially expressed for CHR-P subjects compared to CHR-NP or compared to UC subjects. Comparison of CHR-P to UC subjects using permutation analyses with SAM (Tusher et al, 2001) assuming a false discovery rate of less than 12% identified 30 analytes as differentially expressed {see Table 3). It can thus be expected that of the 30 analytes less than four are false positives. Conducting the same permutation analysis comparing CHR-P to CHR-NP identified only two differentially expressed analytes, as shown in Table 2. These results indicated that distinguishing CHR-P from CHR-NP was much more difficult that distinguishing CHR-P from UC.

Table 3

Results of Permutation Analysis

(Felmeden & Lip, 2005; Page

& Liles, 2013)

Thrombopoietin² P40225 1.25 6.75 Plasma levels regulated by circulating platelets, circulating platelets in schizophrenia have elevated markers of lipid peroxidase damage, and thus may have higher turnover. (Rezin et ah, 2009, Dietrich-Muszalska & Kontek, 2010)

Serum amyloid P- P02743 1.22 6.75 Elevated in acute phase of component inflammation. (Zahedi &

Whitehead, 1989)

Chemokine (c-c P13500 1.22 6.75 See Table 4.

motif) ligand 2¹

Interleukin-15 P40933 1.25 6.75 Pleiotropic, involved in both innate and adaptive immune systems. (Perera et ah, 2012)

Macrophage P14174 0.91 9.67 A pro-inflammatory cytokine migration inhibitory produced by the activated T factor lymphocytes, macrophages, and in the anterior pituitary from arenocorticotropic hormone and thyroid stimulating hormone cells. (Nishino et ah, 1995) Previously described as altered in schizophrenia. (Schwarz et ah, 2013)

Metalloproteinase P01033 1.11 11.35 In addition to inhibiting inhibitor 1 metalloproteinase, has cell growth-promoting activities.

(Visse & Nagase, 2003)

Apolipoprotein D¹ P05090 1.13 11.35 See Table 4.

Apolipoprotein A2 P02652 1.13 11.35 Associates with high density lipoprotein cholesterol. (Tailleu ei al, 2002)

Chemokine (C-C P78556 1.25 11.35 Pro-inflammatory chemokine, motif) ligand 20 increased expression associated with autoimmune disease. (Li, Qi et al, 2013)

Eotaxin'² P51671 1.29 11.35 A chemokine implicated in allergic response, increased expression associated with aging (Villeda et al, 2011) and with use of marijuana. (Femandez-Egea, Scoriels et al, 2013) Serum elevations associated with chronic schizophrenia. (Teixeira et al, 2008)

Matrix P14780 1.17 11.35 A matrix metalloproteinase metalloproteinase 9 and thus involved in proteolysis of extracellular matrix. Associated with inflammation. (Yabluchanskiy et al, 2013) Elevated activity reported in patients with chronic schizophrenia. (Chang e? al, 2011)

Calbindin P05937 1.11 11.35 Involved in regulation of calcium homeostasis. (Bronner, 2001)

Myeloperoxidase P05164 1.18 11.35 An oxidizing enzyme elevated with oxidative stress and inflammation. (Anatoliotakis et al, 2013, Nussbaum et al, 2013) Elevations associated with chronic schizophrenia. (Al-Asmari & Khan, 2013)

Differentially Expressed Analytes, CHR-P vs. CHR-NP*

Matrix P09237 1.25 0 See Table 4.

metalloproteinase 7¹

Apolipoprotein D¹ P05090 1.12 0 See Table 4.

Included in the presently disclosed analyte diagnostic assay selected by the greedy algorithm.

Also included in the blood analyte panel as reported in Schwarz et al, 2011.

Also included in the blood analyte panel as reported in Schwarz et al, 2012a, b, c. ⁴ Different in schizophrenia compared to unaffected comparison subjects in Ramsey et al, 2013.

Plasma Analytes and Psychosis Risk Prediction. An analyte could be chosen up to 125 times with each of the 20 runs (2500 total executions) in the cross-validation procedure. As expected, somewhat different combinations of analytes were chosen every time, but certain analytes were very frequently chosen. The average and quartiles of frequencies are shown in Figure 3. The most confidence in the informativeness of analytes should likely be placed with those most frequently chosen. It was observed, for example, that malondialdehyde-modified low-density lipoprotein was selected in almost all of the 2500 executions of the greedy algorithm. However, after the eighteenth most popular analyte (alpha-2-macroglobulin), the frequency fell by 30%, suggesting a cutoff point and hence a selection of 18 analytes.

The choice of 18 analytes was applied to the full data for all 107 samples, even though the same samples were used in the many 80% subsets to build the classifier. After noting that this is commonly done but is not a best practice, the results might still give an indication of general performance. As shown in Table 4, using the true data the sum of the most frequently chosen 18 analytes gave the highest AUC, but it was only marginally better than using the sum of the most frequently chosen 10 analytes and 5 analytes. Table 4

Area under the Receiver Operating Curve (AUC)

using Various Cut-off Points for Analytes

Best practices of classifier development mandate use of an external test set that is not used to derive the classifier. Only after the classifier is declared is the external test set to be opened and evaluated (see Tropsha, 2010). However, the number of subjects in this study was too small to set aside an external test set. It is known that classifiers generally can find patterns in random data, but the patterns of such pseudo-classification should be weak in some sense compared to patterns from true data. As a feasible alternative method to evaluate whether our findings are chance, exactly the same classifier development process was applied after randomly scrambling all the 107 group memberships (so 35 randomly selected samples became the first type, 40 the second, and 32 the third). The types of the data were randomly permuted and were allocated into bins of 35, 40, and 32 samples (same as UC, CHR-NP, CHR-P). Then, classifiers were built from the pseudo data using the same greedy algorithm and five-fold cross validation process, retaining the sum of the five most frequently selected analytes in every trial. Both true data and 100 trials of pseudo data were thereby used in 101 classifiers with sums of the five most frequently chosen analytes. It terms of AUCs, the true data classifiers had higher performance than all (CHR-NP verus CHR-P) or all but one (UC versus CHR-P) classifier built with random data (and applied to the same random data), suggesting that the true data produced results that are unlikely by chance. Without distributional assumption the p- values were less than 0.01. Noting the skewed distribution of values for the AUC a beta distribution provided a good fit; as shown in Figures 4 A and 4B based on a beta distribution the p-value for random data producing an AUC equal to 0.812 in trie UC versus CHR-P distribution was equal to 0.018, for an AUC equal to 0.823 in the CHR-NP versus CHR-P distribution was equal to 0.000014.

The total frequencies of choices of the first through fifth most frequently chosen analytes (maximum possible frequency was 125) were checked, and summed the top five frequencies (maximum possible = 625). This test reward parsimonious use of data resources, a criterion that is contrary to overfitting. Again, the true classifiers outperformed almost all of the pseudo classifiers (Figure 5). Just comparing the numbers of the two most frequently chosen (which were both "not not" normal at > 5% confidence level) with Student t-test revealed a p- value of less than 7 x 10^"14. The collective frequencies of the first five chosen in classifier versus pseudo-classifier were unquestionably distinct. Again, these results indicated that the data with true labels fundamentally could be classified more accurately than that with random labels.

Additionally, that the potential confounder variables age, sex, socioeconomic status, ancestry, time of blood draw, severity of symptoms, antipsychotic prescription status, and antidepressant prescription status did not impact the relationship of the 18- analyte index to group status using logistic regression was tested. For CHR-P versus CHR-NP, the Wald chi-square estimate for the 18-analyte index in the model containing the index alone was 18.14, (AUC = 0.86) and in the full model including the 18-analyte index and the potential confounder variables was 15.46 (AUC = 0.94). For CHR-P versus UC, the Wald chi-square estimate for the model containing the 18-analyte index alone was 15.73 (AUC = 0.90) and in the full model containing the 18-analyte index and the potential confounders was 14.37 (AUC = 0.91).

Bioinformatic Investigation of Analvte Functions. A summary of functions of the top 18 analytes is shown in Table 5. The summary suggested that oxidative stress and immune and hypothalamic-pituitary system dysregulation might play a role in the development of psychosis. The most frequently chosen analyte, malondialdehyde-low density lipoprotein, measures a lipoprotein damaged by free radicals and is thus a measure of oxidative stress (Del Rio et ah, 2005). Half of the CHR-P subjects had levels greater than a standard deviation from the UC mean, compared to 25% of CHR-NP subjects (Fisher Exact test /j-value = 0.047) and 20% of UC subjects (Fisher Exact test p- value = 0.01 17). Table 5

Bioinformatic Interpretation of Analytes Selected by the Greedy Algorithm using 5-fold

Cross Validation with Re-sampling

axis. (Nadjar et al, 2010) Stimulates the release of pituitary proteins except for prolactin. (Haedo et al, 2009) Induces release of metalloproteinases. (Klein et al, 1997; Vecil et al, 2000) It can induce expression of interleukin-8. (Steude et al, 2002) Reported to be elevated in persons with schizophrenia. (Miller et al, 2011)

Immunoglobulin E P01854 Found only in human, receptors are expressed on mast cells, monocytes, macrophages, and other white blood cells. Classically known to mediate allergic responses. Activation of perivascular (blood brain barrier) mast cells in the hypothalamus by IgE results in HPA axis activation. (Theoharides & Konstantinidou, 2007; Lindsberg et al, 2010)

Uromodulin P07911 Made by the kidney tubules and excreted with urine, as well as by the choroid plexus .(Schuller et al, 1984; Zalc et al, 1984) Uromodulin induces innate immune response(Ratliff, 2005; Weichhart et al, 2005) and stimulates monocytes to release proinflammatory cytokines. (Su et al, 1997)

Transthyretin P02766 Circulating transthyretin tetramer is produced by liver, and monomer produced by choroid plexus. (Redzic & Segal, 2004) Functions to transport thyroid hormone and retinol. Characteristically decreased in acute phase immune response (a negative acute phase reactant), elevation in our sample may indicate increased blood brain barrier/blood CSF barrier permeability; this hypothesis can be tested by looking at transthyretin monomer in blood, which is extremely low in persons with intact BBB and elevated with BBB disruption. (Marchi et al, 2003)

Growth Hormone P01241 Produced by pituitary in response to hypothalamus signaling with growth hormone releasing hormone.

Typically low in acute phase response.

KIT ligand² P21583 Circulating KIT ligand is produced by fibroblasts, endothelial cells, and leptin receptor expressing perivascular stromal cells (Ding et al, 2012; Lennartsson & Ronnstrand, 2012). In the adult KIT ligand is a pleotropic cytokine, and is important for stem cell development, especially hematopoietic stem cells. KIT ligand signaling is important for mast cell responses, including degranulation and cytokine production. KIT Ligand is also associated with dendritic cell activation, promoting release of IL-6. Administration of KIT ligand induces hypothalamic release of adrenocorticotropin (Kovacs et al, 1996). Elevations previously reported in persons with schizophrenia (Schwarz, Guest et al, 2012).

Interleukin-8⁴ P10145 Produced by numerous cells including macrophages and epithelial cells, involved in innate immune response, is an acute phase reactant. Regulates hypothalamic-pituitary response to stress (Rostene et al, 2011). Elevations previously reported in persons with schizophrenia (Miller et al, 2011).

Apolipoprotein D P05090 A lipid-binding molecule involved in transport of hydrophobic molecules (HDL, progesterone, arachadonic acid). Apolipoprotein D is up- regulated with oxidative stress (Ganfornina et al, 2008) . Levels increased in plasma of recent onset schizophrenia (Mahadik et al, 2002), and decreased in the serum of persons with chronic schizophrenia. (Thomas et al, 2001)

Mucin- 16^J Q8WXI7 A marker for ovarian and other cancers, and cardiovascular disease, elevated with inflammatory processes (Hamdy, 2011).

Factor 7²'⁴ P08709 Positive acute phase reactant (inflammation).

Chemokine (C-C P13500 Released by endothelial cells, astrocytes, and motif) ligand 2 microglia. Recruits monocytes, dendritic cells, and

T-lymphocytes to site of inflammation, activates mast cells (Castellani et al, 2010). Elevated via sympathetic system activation in response to social stress (Hanke et al, 2012). Receptors located in several hypothalamic nuclei, including the paraventricular nuclei, a region that integrates neuroendocrine, autonomic, and behavioral reactions to stress (Banisadr et al, 2002; Banisadr et al, 2005; Rostene et al, 2011).

Resistin Q9HD89 The adipokine resistin is an insulin-antagonizing factor that also plays a regulatory role in inflammation, immunity, food intake, and gonadal function. (Rodriguez-Pacheco et al, 2009) In humans, secreted by immune and epithelial cells, increases IL1 beta production, proinflammatory (Miralbell et al, 2012), associated with inflammation but not BMI in obese adolescents (Maggio et al, 2012). Resistin regulates growth hormone (Rodriguez-Pacheco et al, 2009; Rodriguez-Pacheco et al, 2013) and thyroid stimulating hormone (Cinar and Gurlek, 2013) secretion in hypothalamic-pituitary axis

Cortisol²'³ 50-23-7 Hypothalamic dysregulation, inflammation, key hormone of the stress response.

Chemokine (C-C P80075 Produced by monocytes, endothelial cells, microglia motif) ligand 8 (also fibroblasts, epithelial cells), Induced by IL1 beta (among others), modulates mast cells, chemotaxic for monocytes, lymphocytes. Regulates BBB permeability. Alpha-2- P01023 Protease inhibitor, including inhibition of matrix magroglobulin⁴ metalloproteinases, released with blood brain barrier failure by perivascular astrocytes. (Cucullo et al, 2003) Previously reported to be elevated in schizophrenia.

² Also included in the 1 )lood analyte panel as reported in Schwarz et al, 2011.

Also included in the blood analyte panel as reported in Swartz et al, 2012a, b, c.

4Different in schizophrenia compared to unaffected comparison subjects in Ramsey et al, 2013.

⁵SWISS-PROT or CAS Registry.

Elevation of apolipoprotein D, found at higher levels in CHR-P relative to CHR- NP and UC subjects, is also associated with oxidative stress (Ganfornina et al, 2008). In addition, as described in Table 4, most of the chosen analytes are either involved in the inflammatory response or elevated with inflammation. As shown in Table 4, several analytes are related to hormones of the hypothalamic-pituitary axes. These findings led to the hypothesis that dysregulation of inimune-hypothalamic-pituitary interactions could be at play at the emergence of psychosis.

Additional Analytes Identified for Subjects on Medication. Other analytes that could be added to the primary index in special cases of persons not be treated with antipsychotics or antidepressants were also identified. For subjects that were not being treated with an antidepressant, chromogranin A and/or endothelin 1 were identified. For subjects that were not being treated with an antipsychotic, N-(alpha)-acetyltransferase 15, ferritin, and/or alpha 1-anti-chymotrypsin were identified.

Analytes excluded due solely to having a relationship with antipsychotic use were added back, and the greedy algorithm was re-run for subjects who were not prescribed antipsychotics at the time of the blood draw. In addition to the 5-analyte primary index, the addition of one or more of N (alpha)-acetyltransferase 15, ferritin, and alpha 1- antichymotrypsin improved the predictive ability of the index as measured by AUC.

Similarly, analytes excluded due solely to having a relationship with antidepressant use were added back, and the greedy algorithm was re-run for subjects who were not prescribed antidepressants at the time of the blood draw. In addition to the 18-analyte primary index, the addition of one or more of chromogranin A and endothelin 1 improved the predictive ability of the index as measured by AUC. Gender-specific Additional Analytes. Performing separate analyses in males and females indicated that in addition to the 5-analyte primary index, one or more of interleukin 15 and chemokine (c-c motif) ligand 11 in males, and one or more of calbindin 1, transforming growth factor beta, and cytokine (c-c motif) ligand 18 in females improved the predictive ability of the index as measured by AUC. Thus, other analytes that could be added based on the gender of the subject were also identified that could be added to the primary index. For female subjects, calbindin 1, transforming growth factor beta, and/or cytokine (c-c motif) ligand 18 could be added to the primary index. For male subjects, interleukin 15 and/or chemokine (c-c motif) ligand 11 could be added to the primary index.

Discussion of the EXAMPLE

Disclosed herein is a biomarker assay (in some embodiments, a blood biomarker assay) that improved determination of actual psychosis risk in persons experiencing attenuated psychosis risk syndrome. Clinical criteria alone identified persons with a positive prediction of about 30% in two years. The receiver operating characteristic (ROC) for the 18-analyte index shown in Figure 6 indicated that if a sensitivity of 0.6 is accepted, the specificity will be 0.1. Thus, in a population of clinically high risk (CHR+) persons, 70% of persons identified by the test as positive (CHR-P) would be true positives, and 82% true negatives. This cut-off score for the index can be useful for interventions where the risk or cost of treatment is moderately high; of course other cutoff scores with other levels of sensitivity and specificity could also have clinical utility.

The assay included several quality assurance steps, including evaluation of each antibody in a single-plex assay and batch testing to ensure reproducibility over different lots. The intra-assay coefficient of variance, based on native proteins spiked at the low, medium, and high end of the test dynamic range, was less than 0.15 for -90% of the 181 analytes included in the assay. Test-retest reliability of the analyte assay was assessed in one subject; here the correlation of the assay results was very high (r = 0.84 for all 141 analytes).

There are several unique aspects of the presently disclosed methods to develop a blood-based risk predictor for psychosis that warrant comment. First, given that the analyte concentrations varied 1,000,000-fold, a z-score was created for each analyte, and the z-score was based on the average and standard deviation of the unaffected comparison subjects. The z-score normalization preserved the information enabled the creation of an unweighted "index" (all weights = +1). It is proposed that the elimination of weighting with real numbers minimized the problem of overfitting sometimes encountered with more complex algorithms; even logistic regression weights (related to the odds ratio of the predictor) could be vulnerable to overfitting. Overfitting is a recognized problem in statistical modeling. A case in point with the data disclosed herein is that using different 80% subsamples logistic regression created models with AUC of 1.0 each time (perfect classification).

Second, the relevant patient population will frequently be treated with various medications, especially antidepressants and antipsychotics, and these medications could influence the levels of certain analytes. Less than half of patients were prescribed any medication, and the proportion was insignificantly greater in CH -NP as compared to CHR-P subjects. Given that prescription medication use is likely to be common in CHR patients, analytes that showed a possible (trend level) relation to medication use were eliminated. The results presented herein were confirmed in the subjects not treated with medications, increasing confidence that prescribed medications were not an important driver of differences between groups.

Many of the excluded analytes could inform the disease process (for example prolactin; Bicikova et ah, 2011). However, these were not employed in a classifier because prescription of various levels of various antipsychotics and antidepressants could be common in clinical high risk persons.

Based on bioinformatic interpretation of the analytes, it appeared that for a majority of persons with CHR, development of psychosis was related to activation/dysregulation of the immune-hypothalamic-pituitary (IHP) axis {see Figure 7), perhaps mediated by and/or resulting in oxidative stress. This finding is in concert with the substantial literature documenting oxidative damage in persons with schizophrenia (Flatow et ah , 2013). Oxidative stress results in activation of pro-inflammatory processes, and inflammation leads to elevated oxidative stress. Both abnormalities have been reported in persons with schizophrenia (Goldstein et ah , 2007; Nordholm et ah , 2013). In support of this hypothesis, imaging studies have found increases in hypothalamic (Goldstein et ah, 2007) and pituitary (Nordholm et ah, 2013) volumes in persons at elevated risk for psychosis compared to unaffected subjects. In addition, activation of the hypothalamic-pituitary-adrenal axis, as evidenced by elevated salivary Cortisol has also been reported in persons at clinical high risk (CHR) who developed psychosis as compared clinical high risk who did not develop psychosis and unaffected subjects (see Walder et al, 2010; Walker et al, 2010; Walker et al, 2013).

REFERENCES

The references listed below as well as all references cited in the specification including, but not limited to patents, patent application publications, journal articles, and database entries (including but not limited to entries in the GENBANK®,

UniProtKB/SWISS-PROT, and/or Ensembl biosequences databases and/or the CAS

Chemical database, and also including all annotations and references cited therein) are incorporated herein by reference to the extent that they supplement, explain, provide a background for, or teach methodology, techniques, and/or compositions employed herein.

Abbasi et al. (2012) The British MedicalJournal 345:e5900.

Addington et al. (2012) Schizophrenia Research 142(1 -3):77-82.

Ahluwalia et al. (2013) Atherosclerosis 228(2):478-484.

Akiibinu et al (2012) BMC Res Notes 5 : 169.

Al-Asmari & Khan (2013) Hum Exp Toxicol 33: 115-122.

Anatoliotakis et al (2013) Curr Top Med Chem 13:115-138.

Banisadr et al. (2002) J Neurochem 81:257-269.

Banisadr et al. (2005) J Comp Neurol 489:275-292.

Belger et al. (2012) Biological Psychiatry 71(2):94-5.

Bicikova et al. (2011) Neuroendocrinology Letters 32: 141-147.

Bleakley et al. (2013) Canadian Journal of Cardiology 29(2): 147-150.

Bronner (2001) Scientific World Journal 1:919-925.

Burke (2004) Immunobiology 209:51-56.

Cadenhead et al. (2010) Early Intervention in Psychiatry 4(3):220-6.

Cannon et al. (2008) Archives of General Psychiatry 65:28-37.

Carpenter & Tandon (2013) Asian Journal of Psychiatry 6(3):266-268.

Castellan! et al. (2010) J Biol Regul Homeost Agents 24: 1 -6.

Chan et al. (2009) Schizophrenia Bulletin 37(1):177-188.

Chan et al. (2011) International Review oj Neurobiology. 101:95-144.

Chang et al. (2011) Psychiatry Res 187:341-346.

Charmandari et al. (2005) Annu Rev Physiol 67:259-284.

Cinar & Gurlek (2013) Endocr Connect 2:R30-38.

Coffey et al. (1996) Journal of Psychiatric and Mental Health Nursing 3(2):89-93. Cornblatt et al. (2012) Schizophrenia Bulletin 38(6): 1247-1257.

Craddock et al. (2008) Molecular & Cellular Proteomics 7(7): 1204-1213.

Cucullo et al. (2003) Mol Cell Proteomics 2:234-241.

Da Silva et al. (2002) Br J Pharmacol 135:1634-1640.

Del Rio et al. (2005) Nutrition, Metabolism and Cardiovascular Diseasesl 5:316-328.

Diagnostic and Statistical Manual of Mental Disorders. Fifth ed. (2013) American Psychiatric Association Press, Washington, D.C, United States of America.

Dietrich-Muszalska & Kontek (2010) Psychiatry Clin Neurosci 64:469-475.

Ding et al. (2012) Nature 481:457-462.

Dupont et al. (2009) Front Biosci (Landmark Ed) 14: 1337-1361.

Felmeden & Lip (2005) Expert Opin Investig Drugs 14:1319-1336.

Fernandez-Egea et al. (2013) Prog Neuropsychoph.arma.col Biol Psychiatry 46:25-28.

First et al. (2002) Structured Clinical Interview for DSM-IV TR Axis I Disorders, Non- patient Edition (SCID-I/NP). New York: Biometrics Research, New York State Psychiatric Institute, New York, New York, United States of America.

Flatow et al. (2013) Biological Psychiatry 74:400-409.

Fusar-Poli et al (2012) Archives of General Psychiatry 69(3):220-209.

Ganfornina et al. (2008) Aging Cell 7:506-515.

Goh et al. (2013) Expert Review of Cardiovascular Therapy 11(4):425-435.

Goldstein et al. (2007) Biological Psychiatry 61 :935-945.

Golembo-Smith et al. (2012) Schizophrenia Research 139(1-3):129-135.

Green et al. (2009) J Am Acad Child Adolesc Psych 48(11):1060-1068.

Haedo et al. (2009) Horm Res 72:266-27 '4.

Hamdy (2011) Med Sci Monit 17:CR174-179.

Hand (2006) Statistical Science 21:1-34.

Hand (2010) Statistics in Medicine 29(14):1502-10.

Handley & McNeil (1982) Radiology 143:29-36.

Hanke ei fl/. (2012) Brain Behav Immun 26:1150-1159.

Healey ei a/. (2013) Schizophrenia Research 131(1-3): 11-17.

Herberth et al. (2013) Clinical Schizophrenia & Related Psychoses 2013 Feb 8:1-14 [Epub ahead of print].

Hoffmann & Jagla (2002) Int Rev Cytol 213 : 147- 181.

Hope et al. (2009) Bipolar Disord 11(7):726-734. Izmailov et al. (2011) International Review of Neurobiology 101:279-298.

Jun ei a/. (2012) Molecules 17(3):2474-90.

Klein et al. (1997) J Biol Chem 272(22):14188-14192.

Kohavi (1995) Proceedings of the Fourteenth InternationalJoint Conference on Artificial Intelligence 12:1137-1143.

Kovacs et al. (1996) J Neuroimmunol 65(2):133-141.

Lee & Colagiuri (2013) Diabetes Research and Clinical Practice 100(2):285-228.

Lencz et al. (2004) Schizophrenia Research 68:37-48.

Lennartsson & Ronnstrand (2012) Physiol Rev 92(4): 1619-1649.

Levitan et al. (2010) Antioxid Redox Signal 13(1 ) :39-75.

Li et al. (2012) Mol Biosyst 8(10):2664-2671.

Li et al. (2012) PLoS One 7(5):e37195.

Li et al. (2013) PLoS One 8(5):e64277.

Lindsberg et al. (2010) JCereb Blood Flow Metab 30(4):689-702.

Liu et al. (2005) BMC Bioinformatics. 6:76.

Maggio et al. (2012) J Pediatr 161(6):1022-1027.

Mahadik et al. (2002) Schizophrenia Research 58(l):55-62.

Maliner-Stratton et al. (2001) Neoplasia 3(6):509-520.

Marchi et al. (2003). J Neurosci 23(5): 1949-1955.

Miller et al. (2002) The American Journal of Psychiatry 159(5):863-5.

Miller et al. (2003) Schizophrenia Bulletin 29(4):703-715.

Miller et al. (2011) Biological Psychiatry 70(7):663-671.

Miralbell et al. (2012) Neurobiol Aging 33(5):1003 el009-1017.

Moons et al. (2012) Heart (British Cardiac Society) 98(9):683-90.

Moons et al. (2012) Heart (British Cardiac Society) 98(9):691-8.

Nadjar et al. (2010) NeurobiolDis 39(3):393-401.

Nagai et al (2013) Schizophrenia Research 150(2-3):547-554.

Nishino et al. (1995) Mol Med l(7):781-788.

Nordholm et al. (2013) Psychoneuroendocrinology 38, 2394-2404.

Nussbaum et al. (2013) Antioxid Redox Signal 18(6):692-713.

Page & Liles (2013) Virulence 4(6):507-516.

PCT International Patent Application Publication Nos. WO 2003/025224; WO 2013/186562. Perera et al. (2012) Microbes Infect 14(3):247-261.

Perez et al. (2013) Biological Psychiatry 2013 [Epub ahead of print],

Perkins et al. (2005) The American Journal of Psychiatry 162: 1785-1804.

Ramsey et al. (2013) PLoS One 8(1 l):e78729.

Ratliff (2005) J Urol 174(3): 1150.

Redzic & Segal (2004) Adv Drug Deliv Rev 56(12): 1695-1716.

Rezin et al. (2009) Neurochem Res 34(6): 1021-1029.

Rodriguez-Pacheco et al. (2009) Endocrinology 150(10):4643-4652.

Rodriguez-Pacheco et al. (2013) Mediators Inflamm 2013:479739.

Rosier et al. (2010) Arthritis Rheum 62(3):815-825.

Rostene et al. (201 1 ) Front Neuroendocrinal 32(1): 10-24.

Schuller et al. (1984) Eur Neurol 23(2):65-72.

Schultze-Lutter et al. (2013) Psychopathology 46(2):75-87.

Schwarz et al. (2010) Biomarker Insights 5:39-47.

Schwarz et al. (2011) Molecular Psychiatry 201 1 : 1-9.

Schwarz et al. (2012a) Translational Psychiatry 2:e82.

Schwarz et al. (2012b) The World Journal of Biological Psychiatry 13(8): 627-632.

Schwarz et al. (2012c) Molecular Psychiatry 17(5):494-502.

Schwarz et al. (2013) Schizophr Bull, doi: 10.1093/schbul/sbtl 05.

Seidman et al. (2010) Archives of General Psychiatry 67(6):578-588.

Shao et al. (2006) Ann N Y Acad Sci 1072: 194-209.

Shin et al. (2009) Biological Psychiatry 65(12): 1071 -8.

Sowa et al. (2010) J Neurosci 30(31): 10282- 10293.

Steiner ei a/. (2010) Molecular Psychiatry. 15(l):3-4.

Steude et al. (2002) J Invest Dermatol 119(6): 1254-1260.

Su et al. (1997) J Immunol 158(7):3449-3456.

Suleiman et al. (2013) Crit Rev Oncol Hematol. 88:637-654.

Tailleux et al. (2002) Atherosclerosis 164(1): 1-13.

Takahashi et al. (2009) Archives of General Psychiatry 66(4):366-376.

Teixeira et al. (2008) Prog Neuropsychopharmacol Biol Psychiatry) 32(3):710-714.

Theoharides & Konstantinidou (2007) Front Biosci 12: 1615-1628.

Thomas et al. (2001) Proc Natl Acad Sci USA 98(7):4066-4071. Tomasik et al. (2012) European Archives of Psychiatry and Clinical Neuroscience 262 Suppl 2:S79-83.

Tropsha (2010) Molecular Informatics 29:476-488.

Tsang et al. (2006) Journal ofProteome Research 5(4):756-760.

Tsuang et al. (2013) Schizophrenia Research, doi: 10.1016/j.schres.2013.05.004.

Tusker et al. (2001) Proc NatAcaad Sci USA 98:5116-5121.

U.S. Patent Application Publication No. US 2014/0200151.

U.S. Patent No. 8,492,418.

Vecil et al. (2000) JNeurosci Res 61(2):212-224.

Villeda et al (2011) Nature 477(7362):90-94.

Visse & Nagase (2003) Circ Res 92(8):827-839.

Walder et al. (2010) Psychiatric Genetics 20166-170.

Walker et al. (2010) Journal of Abnormal Psychology 119:401-408.

Walker et al. (2013) Biological Psychiatry 74:410-417.

Wang et al. (2009) Glia 57(l):24-38.

Weichhart et al. (2005) Wien Klin Wochenschr 117(9-10):316-322.

Witthaus et al. (2008) Schizophrenia Research 102(1-3):141-149.

Wu et al. (2013) Prog Neuropsychophannacol Biol Psychiatry 46:200-206.

Yabluchanskiy et al. (2013) Physiology (Bethesda) 28(6):391-403.

Zahedi & Whitehead (1989) J Immunol 143(9): 2880-2886.

Zalc et al. (1984) 5raz^'ra jtes 291(l):182-187.

It will be understood that various details of the presently disclosed subject matter may be changed without departing from the scope of the present subject matter. Furthermore, the foregoing description is for the purpose of illustration only, and not for the purpose of limitation.

Claims

CLAIMS What is claimed is:

1. A method of identifying a risk of developing a psychotic disorder in a subject, the method comprising:

(a) isolating and/or providing a biological sample comprising serum, plasma, urine, and/or saliva from a subject who meets clinical criteria for elevated risk of psychosis;

(b) quantifying a level of expression in the biological sample of a plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample wherein the gene products or other endogenous or exogenous substances, or the mixture thereof, are selected from the group consisting of malondialdehyde- modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E;

(c) combining the quantified levels of expression from step (b) to create a summary measure of expression; and

(d) comparing the summary measure of expression for the subject to one or more standards, wherein the one or more standards are selected from the group consisting of:

(i) summary measures of expression in subjects with elevated risk for psychosis who did not develop psychosis;

(ii) summary measures of expression in subjects with elevated risk for psychosis who did develop psychosis; and

(iii) summary measures of expression in unaffected subjects;

wherein the comprising step identifies a risk of developing a psychotic disorder in the subject.

2. A method of identifying whether or not a subject is developing a psychotic disorder, the method comprising:

(b) quantifying levels of expression in the biological sample for a plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample, wherein the gene products or other endogenous or exogenous substances are malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E;

(c) combining the quantified levels of expression from step (b) to create a summary measure of expression;

(iii) summary measures of expression in unaffected subjects;

(e) repeating steps (a)-(d) at a second, later time point; and

(f) assessing whether the result of step (d) as determined using the serum and/or plasma isolated at the second time point as compared to the result of step (d) as determined using the serum and/or plasma isolated at the first time point is indicative of the subject developing a psychotic disorder.

The method of any of the previous claims, wherein at least one of the one or more standards comprises a summary measure that is normalized with respect to healthy subjects of similar age, sex, ancestry, time of day that the sample is drawn, and/or whether taking prescribed medications as the subject.

The method of any of the previous claims, wherein the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in biological samples further comprises one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, <¾- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2. The method of any of the previous claims, wherein the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in biological samples further comprises one or more of beta 2 transferrin, prostaglandin D synthase (beta trace protein), adrenocorticotropin releasing hormone, and insulin-like growth factor, or substances associated therewith.

The method of any of the previous claims, wherein the subject is not being treated with an antidepressant, and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in biological samples further comprises cliromogranin A, endothelin I, or both.

The method of any of the previous claims, wherein the subject is not being treated with an antipsychotic, and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in biological samples further comprises one or more of N-(alpha)-acetyltransferase 15, ferritin, and alpha 1-anti-chymotrypsin.

The method of any of the previous claims, wherein the subject is a female and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in biological samples further comprises one or more of calbindin 1, transforming growth factor beta, and cytokine (c-c motif) ligand 18.

The method of any of claims 1-7, wherein the subject is a male and the plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in a biological sample further comprises one or more of interleukin 15 and chemokine (c-c motif) ligand 11.

The method of any of the previous claims, wherein the subject has or is at risk for developing Attenuated Psychosis Syndrome (APS).

The method of any of the preceding claims, wherein steps (a)-(d) are performed at least two different time points, and the results of the comparing step (d) from the at least two different time points is indicative of the subject developing a psychotic disorder. The method of any of the preceding claims, further comprising administering to the subject an antipsychotic medication if the subject is determined to be at risk of developing a psychotic disorder or is determined to be developing a psychotic disorder.

The method of any of the preceding claims, wherein the quantifying comprises employing a plurality of antibodies, at least one of which binds to each of the gene products or other endogenous or exogenous substances with sufficient specificity to allow for quantification of the gene products or other endogenous or exogenous substances in the biological sample.

The method of any of the preceding claims, wherein the summary measure of expression for the subject is calculated by determining a z-score for each analyte assayed, wherein the z-score is based on the average and standard deviation of the unaffected comparison subjects, and summing the individual calculated z-scores.

The method of any of the preceding claims, wherein the subject is a female and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise:

(1) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, hiterleukin-lB, matrix metalloproteinase 7, and immunoglobulin E; or

(2) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2; or

(3) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of calbindin 1, transforming growth factor beta, and cytokine (c-c motif) ligand 18; or

(4) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of calbindin 1 , transforming growth factor beta, and cytokine (c-c motif) ligand 18.

The method of any of the preceding claims, wherein the subject is a male and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise:

(1) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Inter] eukin- IB, matrix metalloproteinase 7, and immunoglobulin E; or

(2) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-1B, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2; or

(3) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interl eukin- IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of interleukin 15 and chemokine (c-c motif) ligand 11 ; or

(4) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin- IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of interleukin 15 and chemokine (c-c motif) ligand 11.

The method of any of the preceding claims, wherein the subject is not being treated with an antidepressant and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise:

(1) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E; or

(3) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of chromogranin A, endothelin 1, or both; or

(4) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of chromogranin A, endothelin 1, or both.

The method of any of the preceding claims, wherein the subject is not being treated with an antipsychotic and the plurality of gene products or other endogenous or exogenous substances for which level of expressions are quantified comprise:

(3) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of N-(alpha)-acetyltransferase 15, ferritin, and alpha 1- anti-chymotrypsin; or

(4) malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a2- macroglobulin, mucin- 16, and chemokine (c-c motif) ligand 2, and also one or more of N-(alpha)-acetyltransf erase 15, ferritin, and alpha 1-anti- chymotrypsin.

A kit comprising a plurality of reagents for detecting the presence of and/or quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances, or mixtures thereof, wherein the gene products or other endogenous or exogenous substances, or the mixture thereof, are selected from the group consisting of malondialdehyde- modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and optionally further include one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, <¾-macroglobulin, mucin-16, chemokine (c-c motif) ligand 2, of beta 2 transferrin, prostaglandin D synthase (beta trace protein), adrenocorticotropin releasing hormone, insulin-like growth factor, chromogranin A, endothelin 1, N-(alpha)-acetyltransferase 15, ferritin, alpha 1- anti-chymotrypsin, calbindin 1, transforming growth factor beta, cytokine (c-c motif) ligand 18, interleukin 15, and chemokine (c-c motif) ligand 11.

The kit of claim 19, wherein the reagents comprise antibodies that bind to malondialdehyde-modified low density lipoprotein, thyroid stimulating hormone, Interleukin-IB, matrix metalloproteinase 7, and immunoglobulin E, and optionally further comprise antibodies that binds to one or more of transthyretin (either total or low molecular weight), uromodulin, growth hormone, KIT ligand, IL-8, apolipoproptein D, chemokine (c-c motif) ligand 8, Factor 7, Cortisol, resistin, a₂- macroglobulin, mucin- 16, chemokine (c-c motif) ligand 2, of beta 2 transferrin, prostaglandin D synthase (beta trace protein), adrenocorticotropin releasing hormone, insulin-like growth factor, chromogranin A, endothelin 1, N-(alpha)- acetyltransferase 15, ferritin, alpha 1-anti-chymotrypsin, calbindin 1, transforming growth factor beta, cytokine (c-c motif) ligand 18, interleukin 15, and chemokine (c-c motif) ligand 11.

The kit of claim 19 or claim 20, wherein the reagents for detecting the presence of and quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances are detectably labeled.

The kit of any one of claims 19-21, wherein the kits further comprise compounds that permit detection of the detectably labeled reagents.

The kit of any one of claims 19-22, wherein one or more of the reagents provided in the kit is labeled with a different detectable label, allowing for simultaneous detection and quantification of multiple gene products or other endogenous or exogenous substances in biological samples.

The kit of claim 19, wherein the reagents for detecting the presence of and quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances are antibodies, optionally monoclonal antibodies, and optionally fragments of antibodies that comprise paratopes that bind to one or more of the gene products or other endogenous or exogenous substances.

The kit of claim 19 or claim 24, wherein one or more of the antibodies are affixed to a solid support.

The kit of claim 19, wherein the reagents for detecting the presence of and quantifying levels of expression in biological samples of pluralities of gene products or other endogenous or exogenous substances are oligonucleotide probes that are specific for one or more of the gene products or other endogenous or exogenous substances. The kit of claim 26, wherein one or more of the oligonucleotide probes comprises a nucleotide sequence designed to bind to a cDNAs derived from a gene product but not bind to genomic DNA by virtue of having sequences that are derived from from different exons of the gene products.

A non-transitory computer readable medium having stored thereon computer executable instructions that when executed by a processor of a computer controls the computer to perform steps comprising: quantifying a level of expression in the biological sample of a plurality of gene products or other endogenous or exogenous substances, or a mixture thereof, normally found in the biological sample wherein the gene products or other endogenous or exogenous substances, or the mixture thereof, are selected from the group consisting of malondialdehyde- modified low density lipoprotein, thyroid stimulating hormone, Inteiieukin-1B, matrix metalloproteinase 7, and immunoglobulin E; combining the quantified levels of expression to create a summary measure of expression; and comparing the summary measure of expression for the subject to one or more standards, wherein the one or more standards are selected from the group consisting of: (i) summary measures of expression in subjects with elevated risk for psychosis who did not develop psychosis;

(ϋ) summary measures of expression in subjects with elevated risk for psychosis who did develop psychosis; and

(iii) summary measures of expression in unaffected subjects.