Biomarkers for bipolar disorder

Abstract

Biomarkers for bipolar disorder include Cer(d18:0 / 24:1), TG(18:1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17:1_36:3), SM (OH) C14:1, SM C20:2, PC aa C42:0, and HexCer(d18:1 / 22:0). Levels of the biomarkers can be measured in samples taken from patients for use in methods of diagnosis of bipolar disorder or a predisposition to bipolar disorder, or for prognosing the development of bipolar disorder, or for monitoring efficacy of a therapy in an individual having, suspected of having, or being predisposed to bipolar disorder.

Description

[0001] BIOMARKERS FOR BIPOLAR DISORDER

[0002] TECHNICAL FIELD

[0003] This invention relates to biomarkers and methods of diagnosing, prognosing or monitoring bipolar disorder, or a predisposition thereto, and to methods of treatment of bipolar disorder.

[0004] BACKGROUND

[0005] Major depressive disorder (MDD) and bipolar disorder (BD) are complex and debilitating conditions affecting 16.2% and 1% (GBD 2019 (2020)) of the population worldwide, respectively (Merikangas et al. (2011); Kupfer et al. (2012)), and are among the leading contributors to the global burden of diseases (GBD 2019 (2020)). Core symptoms of MDD include a pervasive and persistent disturbance of mood and loss of interest / pleasure in most daily activities (Otte et al. (2016)). Individuals can also experience impaired concentration and indecisiveness, as well as fatigue or low energy, disturbances to sleep and appetite, headaches, muscle tension, and general symptoms of pain (American Psychiatric Association (2013)). BD, on the other hand, is typically characterized by intermittent depressive and manic (BDI) or hypomanic (BDII) episodes. BD can also include psychotic symptoms such as delusions and hallucinations (Paykel et al. (2008)). Disease onset is commonly in late adolescence or early adulthood, affecting men and women equally. When compared to the general population, BD is a life shortening condition.

[0006] While depressive episodes in BD may be indistinguishable from those in MDD, (hypo)manic episodes can include elevated levels of energy, euphoric mood, irritability, and hypersexuality, as well as a reduced need for sleep (Einat, 2007). To date, diagnosing MDD and BD requires a comprehensive collection of symptom-and patient-level data, as well as information on family history, course of illness, and prior treatment response. The World Health Organization World Mental Health Composite International Diagnostic Interview (CIDI (Kessler & Ustun (2004))) can be a useful tool for the assessment and differential diagnosis of MDD and BD. It also captures subthreshold forms of elevated mood, including subthreshold BD and MDD with subthreshold BD, with individuals presenting with fewer and / or a shorter duration of (hypo)manic symptoms. Both MDD and BD typically start early in life and are associated with severe functional impairments, high morbidity and mortality, including premature death due to suicide (Bourne et al. (2013); Passos et al. (2016); Osby et al. (2016)).

[0007] The economic burden of MDD and BD is also substantial; in England, direct and indirect annual costs are estimated at £7.46 billion (US $9.85 billion) for MDD and £5.25 billion (US $6.93 billion) for BD (McCrone et al. (2008)). BD has a substantial impact on the European Union (EU) population and economy. In 2010, an estimated 3 million people (0.9% of the population) within the EU had been diagnosed with BD, amounting to a total cost of €21.5 billion, with the majority of costs (€18.0 billion; 83.7%) being indirect such as lost productivity (e.g. sick leave).

[0008] In spite of their prevalence and negative prognosis, the recognition and diagnosis of BD and MDD presents a significant challenge, particularly in the primary care setting. For instance, research has shown that general practitioners (GPs) initially misdiagnose 50% of MDD patients (Mitchell et al. (2009)). Short consultation times coupled with the difficulties associated with diagnosing MDD, where any two individuals may have no symptoms in common (Olbert et al. (2014); Fried et al., 2014), means that many are not receiving the support they need. This is a particular issue, given that the vast majority of patients with MDD receive treatment and care solely in the primary care setting.

[0009] Diagnosis of BD is generally based upon clinical interviews endeavouring to identify BD mood symptoms and patterns. In most cases, the depressive symptoms at the initial presentation of BD overlap with symptoms of MDD, whilst manic symptoms overlap with symptoms observed in schizophrenia. This overlap of symptoms frequently results in BD being misdiagnosed leading to long delays between the onset of initial symptoms until correct diagnosis.

[0010] In the case of BD, almost 40% of individuals are initially misdiagnosed with MDD, despite having experienced a prior manic or hypomanic episode (Ghaemi et al. (2000); Hirschfeld et al. (2003)). This results in an estimated delay between the onset of bipolar symptoms and their adequate management of 5.9 years (Scott et al. (2022)). This is due, in part, to the fact that the majority of individuals with BD seek help when they are experiencing depressive symptoms as opposed to when they are in a (hypo)manic state. Furthermore, in most instances, depressive episodes precede a first (hypo)manic episode, and awareness of one's (hypo)manic symptoms is relatively low (Regeer et al. (2015)). This poses a significant problem for the diagnosis, treatment, and management of BD, with these individuals likely to be treated with antidepressant monotherapy, which is frequently ineffective in treating bipolar depression. Critically, the use of antidepressants without a concurrent mood stabilizer can trigger and exacerbate a (hypo)manic episode, resulting in prolonged suffering and, in some cases, suicide (Bowden (2005)).

[0011] Taken together, the careful evaluation and management of all patients presenting with depressive symptoms is warranted. Even in the absence of a (hypo)manic episode, individuals diagnosed with MDD should be closely monitored and managed. This is because symptoms of MDD are frequently the initial presentation of BD, with factors including greater depression severity, recurrent MDD, and psychotic symptoms associated with a later BD diagnosis (Holma et al. (2008)). Other, patient-level risk factors that are indicative of the disorder comprise an earlier age of onset of depressive symptoms, being white, living alone, not being married, and being unemployed (Hirschfeld et al. (2005)).

[0012] The collection of extensive symptom- and patient-level data may prove difficult in the primary care setting, where time is a luxury. The subjective nature of psychiatric evaluations based on clinicians' impressions of patient self-reported symptoms can also be problematic. Digital technologies may offer an innovative, time-and cost- effective alternative to conventional, interview-based methods. A comprehensive and careful appraisal of the characteristics of individuals presenting with depressive symptoms is likely to improve biological disease understanding, facilitate patient stratification, and allow for personalized treatment plans and strategies.

[0013] Despite the established clinical need for an objective test for the diagnosis of BD to be routinely used in conjunction with clinical interviews (Schwartz & Bahn (2008);

[0014] Schwarz et al. (2009); Bragazzi (2013)) extensive research into neuroimaging based biomarkers (Atluri et al. (2013)) and genetic risk factors (e.g. CACNAIC, ODZ4, and NCAN (Craddock & Sklar (2013)), has as yet not resulted in a diagnostic test for routine clinical use. Biomarker profiling offers a promising approach to overcoming the existing challenges and enabling earlier and more objective differential diagnosis of mood disorders (Brunkhorst-Kanaan et al. (2019); Tkachev et al. (2023); de Kluiver et a / . (2023)). A promising approach involves the identification of robust differential biomarkers in peripheral tissues that could support the clinical decision-making process. Several studies have attempted to distinguish BD from MDD using high-dimensional biological data, including those at the genetic (Liebers et al. (2021)), gene expression (Powell (2014); Chen (2020)), protein (Kittel-Schneider et a / . (2020); Tomasik et a / . (2021); Idemoto et a / . (2021)), metabolite (Brunkhorst-Kanaan et a / . (2019)), and microbiome level (Li et al. (2022)). Although the findings have indicated a potential utility of biomarker profiling for differentiating the disorders, several limitations have hindered their clinical applicability.

[0015] Key unaddressed challenges include identifying disease biomarkers in patients with established BD where the confounding effect of mood-stabilizing medication could not be eliminated, using samples from patients with diverging symptom polarities and other differing characteristics that are not accounted for, and the absence of independent validation cohorts.

[0016] Therefore, there is a need to develop an objective test, in particular a blood-based molecular biomarker test, for identification of bipolar disorder during the early stages of the disease. In particular, there is a need to develop a test that can distinguish BD from MDD during episodes of low mood.

[0017] The applicant has previously investigated an approach based on measuring levels of a panel of different proteins in a biological sample obtained from patients in order to diagnose or monitor bipolar disorder, or a predisposition thereto (WO 2016 / 1135750, the content of which is hereby incorporated by reference).

[0018] The present invention seeks to provide an improved method of diagnosing or prognosing bipolar disorder or a predisposition thereto. SUMMARY OF THE INVENTION

[0019] According to an aspect of the present invention, there is provided use of at least one of: Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14:l, SM C20:2, PC aa C42:0, and / or HexCer(dl8: 1 / 22:0), as a biomarker for the diagnosis of bipolar disorder or a predisposition to bipolar disorder, or for prognosing the development of bipolar disorder, or for monitoring efficacy of a therapy in an individual having, suspected of having, or being predisposed to bipolar disorder.

[0020] According to another aspect of the present invention, there is provided a method of diagnosing bipolar disorder or a predisposition thereto or prognosing the development of bipolar disorder, the method including: (a) measuring the amount of at least one of the following biomarkers in a biological sample from an individual: Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0, and / or HexCer(dl8: 1 / 22:0); and (b) comparing the amount of the biomarker in the individual to a threshold value or to the amount of the biomarker in a control; wherein a difference in the amount of the biomarker is indicative of bipolar disorder or a predisposition thereto.

[0021] According to another aspect of the present invention, there is provided a method of monitoring efficacy of a therapy in an individual having, suspected of having, or being predisposed to bipolar disorder, the method including: measuring the amount of at least one of the following biomarkers in a biological sample from the individual: Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0, and / or HexCer(dl8: 1 / 22:0), wherein: a) the method is carried out on samples taken on two or more occasions from the individual and the amount of the biomarker present in the two or more samples is compared; and / or b) the amount of the biomarker present in the sample from the individual is compared with a threshold value or to one or more controls.

[0022] According to another aspect of the present invention, there is provided a method of treating a bipolar disorder patient including: (a) measuring the amount of at least one of the following biomarkers in a biological sample from said patient: Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0, and / or HexCer(dl8: 1 / 22:0); (b) comparing the amount of the biomarker in the patient to a threshold value or to the amount of the biomarker in a control; wherein a difference in the amount of the biomarker is indicative of bipolar disorder or a predisposition thereto; and (c) prescribing and / or administering a bipolar disorder medicament to said patient.

[0023] The method or use may include measuring the amount of a biomarker selected from the group comprising Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3- IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0, and HexCer(dl8: 1 / 22:0).

[0024] The method or use may include measuring the amount of a biomarker selected from the group consisting of Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3- IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0, and HexCer(dl8: 1 / 22:0).

[0025] Abbreviations: 3-IAA - indoleacetic acid; 3-IPA - indolepropionic acid; Asn - asparagine; Cer - ceramide; GCA - glycocholic acid; HexCer - hexosylceramide; Lac - lactic acid; PC - phosphatidylcholine; SM - sphingomyelin; SM (OH) - hydroxysphingomyelin; t4-OH-Pro - trans-4-hydroxyproline; TG - triacylglyceride; TMAO - trimethylamine N-oxide; TrpBetaine - tryptophan betaine.

[0026] The method of monitoring efficacy of a therapy may include comparing the amount of the biomarker in a sample obtained from the individual with the amount present in one or more samples taken from the individual prior to commencement of therapy, and / or one or more samples taken from the individual at an earlier stage of therapy.

[0027] In a method of monitoring efficacy of a therapy, samples may be taken prior to and / or during and / or following therapy for bipolar disorder.

[0028] The method of treating a bipolar disorder patient may include prescribing and / or administering a mood stabilizing drug such as lithium, an anticonvulsant medication such as valproate, carbamazepine or lamotrigine, an antipsychotic medicine such as haloperidol, olanzapine, quetiapine, aripiprazole or risperidone. In some embodiments, treatment may include prescribing and / or administering a selective serotonin reuptake inhibitor such as citalopram, escitalopram, fluoxetine, or sertraline. In some embodiments, treatment may include prescribing and / or administering a serotonin-norepinephrine reuptake inhibitor such as venlafaxine or duloxetine. In some embodiments, treatment may include prescribing and / or administering a norepinephrine and dopamine reuptake inhibitor, such as bupropion. In some embodiments, treatment may include prescribing and / or administering a tetracyclic antidepressant such as amoxapine, maprotiline, mazindol or mirtazapine. In some embodiments, treatment may include prescribing and / or administering a tricyclic antidepressant such as amitriptyline, imipramine, or nortriptyline. In some embodiments, treatment may include prescribing and / or administering a monoamine oxidase inhibitor such as selegiline, isocarboxazid, phenelzine or tranylcypromine.

[0029] The method or use may include comparing the amount of the biomarker in a sample from the individual with the amount of the biomarker present in a sample from a subject not having bipolar disorder.

[0030] The method or use may include measuring the amount of at least one of the following biomarkers in a biological sample from an individual Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, and / or GCA.

[0031] The method or use may include measuring the amount of a biomarker selected from the group comprising Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3- IAA, Lac, PC aa C32:3, and GCA.

[0032] The method or use may include measuring the amount of a biomarker selected from the group consisting of Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3- IAA, Lac, PC aa C32:3, GCA.

[0033] The method or use may include measuring the amount of at least Cer(dl8:0 / 24: 1) and / or 3-IAA. The method or use may include measuring the amount of at least Cer(dl8:0 / 24: 1). The method or use may include measuring the amount of at least 3-IAA. The method or use may include measuring the amount of at least Asn. The method or use may include measuring the amount of at least 3-IPA. The method or use may include measuring the amount of at least t4-OH-Pro.

[0034] The method or use may include measuring at least two of the biomarkers, at least three of the biomarkers, at least four of the biomarkers, at least five of the biomarkers, at least six of the biomarkers, at least seven of the biomarkers, at least eight of the biomarkers, at least nine of the biomarkers, at least 10 of the biomarkers, at least 11 of the biomarkers, at least 12 of the biomarkers, at least 13 of the biomarkers, at least 14 of the biomarkers, at least 15 of the biomarkers, at least 16 of the biomarkers, or all 17 of the biomarkers.

[0035] In an embodiment, the method or use includes measuring the amount of Cer(dl8:0 / 24: l) and at least one of: TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), PC aa C42:0, HexCer(dl8: 1 / 22:0).

[0036] In an embodiment, the method or use includes measuring the amount of Cer(dl8:0 / 24: l) and at least one of: TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA.

[0037] In an embodiment, the method or use involves measuring only biomarkers selected from Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0, and HexCer(dl8: l / 22:0).

[0038] In an embodiment, the method or use involves measuring only biomarkers selected from Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA.

[0039] The method or use may include calculating an overall probability of the individual having bipolar disorder based on probabilities determined using each biomarker measured, and comparing the overall probability of the individual having bipolar disorder with a threshold probability or to a probability determined from a control. The method or use may include measuring at least Cer(dl8:0 / 24: 1) and

[0040] TG(18: 1_38:5); at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5) and TrpBetaine; at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine and PC ae C36:5; at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5 and 3-IAA; at least Cer(dl8:0 / 24: l), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA and Lac; or at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac and PC aa C32:3.

[0041] The method or use may include measuring at least or only Cer(dl8:0 / 24: 1),

[0042] TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0 and HexCer(dl8: 1 / 22:0).

[0043] In a preferred embodiment, the method or use includes measuring at least or only Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3 and GCA.

[0044] The individual may be suffering with or may have suffered with low mood. The individual may have previously been diagnosed with a mood disorder, such as major depressive disorder. The method or use may be for differential diagnosis of bipolar disorder in an individual presenting with low mood. The individual may have previously been diagnosed with major depressive disorder.

[0045] The biological sample may be blood, which may be in the form of a dried blood spot.

[0046] The quantifying may be performed by measuring the concentration of the biomarker in the or each sample.

[0047] Where the individual is diagnosed as having bipolar disorder or being at risk of bipolar disorder, the method or use may include the step of administering a bipolar disorder medicament to the individual.

[0048] According to another aspect of the present invention, there is provided a method of treating a bipolar disorder patient identified by the use of a biomarker as specified above or using a method as specified above, the method of treating including prescribing and / or administering a bipolar disorder medicament to said patient.

[0049] According to another aspect of the present invention, there is provided a method of treating a bipolar disorder patient, including: administering a bipolar disorder medicament to a patient identified as having or being predisposed to bipolar disorder, wherein the patient has an elevated level of at least one of the following biomarkers: 3-IAA, Asn, Cer(dl8:0 / 24: 1), GCA, Lac, PC aa C42:0, PC ae C36:5, t4-OH-Pro, or TMAO; and / or a decreased level of at least one of the following biomarkers: 3-IPA, HexCer(d 18: 1 / 22:0), PC aa C32:3, SM (OH) C14: l, SM C20:2, TG(17: 1_36:3), TG(18: 1_38:5) and TrpBetaine.

[0050] According to another aspect of the present invention, there is provided a method of diagnosing bipolar disorder or a predisposition thereto or prognosing the development of bipolar disorder, the method including: (a) measuring the amount of the following biomarkers in a biological sample from an individual: Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, and GCA; and (b) comparing the amount of the biomarker in the individual to a threshold value or to the amount of the biomarker in a control; wherein an elevated level of the following biomarkers: 3-IAA, Cer(dl8:0 / 24: 1), GCA, Lac, and PC ae C36:5 is indicative of bipolar disorder or a predisposition thereto; and / or a decreased level of the following biomarkers: PC aa C32:3, TG(18: 1_38:5) and TrpBetaine is indicative of bipolar disorder or a predisposition thereto.

[0051] According to another aspect of the present invention, there is provided a method of diagnosing bipolar disorder or a predisposition thereto or prognosing the development of bipolar disorder, the method including: (a) measuring the amount of the following biomarkers in a biological sample from an individual: Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0, and / or HexCer(dl8: 1 / 22:0); and (b) comparing the amount of the biomarker in the individual to a threshold value or to the amount of the biomarker in a control; wherein an elevated level of the following biomarkers: 3-IAA, Asn, Cer(dl8:0 / 24: 1), GCA, Lac, PC aa C42:0, PC ae C36:5, t4-OH-Pro, or TMAO is indicative of bipolar disorder or a predisposition thereto; and / or a decreased level of the following biomarkers: 3-IPA, HexCer(dl8: 1 / 22:0), PC aa C32:3, SM (OH) C14: l, SM C20:2,

[0052] TG(17: 1_36:3), TG(18: 1_38:5) and TrpBetaine is indicative of bipolar disorder or a predisposition thereto.

[0053] A kit for performing a method as specified above may include a questionnaire for use in diagnosing a patient with BD. The questionnaire may be used to support the results obtained by measuring the biomarkers or to help determine the severity of BD.

[0054] LIST OF DRAWINGS

[0055] Embodiments of the present invention are described below, by way of example only, with reference to the accompanying drawings, in which:

[0056] Figure 1 is a schematic drawing showing the steps of an embodiment of a method of the invention;

[0057] Figures 2 to 18 are plots illustrating the relationship between biomarker concentrations and the likelihood of a diagnosis of BD;

[0058] Figures 19 and 20 are graphs showing the cross-validation and validation of the biomarker-based model;

[0059] Figure 21 illustrates the relative importance of the different biomarkers identified;

[0060] Figures 22 to 38 show scaled biomarker concentrations in the BD and MDD groups;

[0061] Figure 39 shows SHAP analysis of the top biomarkers;

[0062] Figure 40 shows cumulative contribution of compound classes to the overall accuracy of the biomarker-based model;

[0063] Figure 41 illustrates improvement of performance of diagnostic models based on demographic and symptom information when biomarker readouts were combined with patient-reported data;

[0064] Figure 42 shows the relative contribution of patient-reported symptom and demographic features and biomarker data to diagnostic accuracy;

[0065] Figure 43 shows biomarker importance in individual models;

[0066] Figures 44 to 46 show the results of decision curve analysis demonstrating that at relevant diagnostic thresholds, the inclusion of biomarkers may result in the additional identification of BD patients; Figure 47 shows that the identified biomarkers correlated primarily with lifetime manic symptoms and psychiatric history;

[0067] Figure 48 shows an example of a biomarker-based extreme gradient-boosted tree diagnostic model;

[0068] Figure 49 shows the breakdown of a biomarker-based diagnostic prediction for a bipolar disorder patient; and

[0069] Figure 50 shows the breakdown of a biomarker-based diagnostic prediction for a major depressive disorder patient.

[0070] DESCRIPTION

[0071] The results provided herein describe an extensive study which demonstrates for the first time the potential utility of the biomarker panels of embodiments of the invention as a blood-based diagnostic test for BD at the early stages of the disease. An early and accurate diagnosis has the potential to improve patient quality of life, prevent unnecessary suffering, enhance treatment outcomes, and reduce healthcare costs.

[0072] References to "bipolar disorder" as used herein refer to a mental disorder characterised by remitting and relapsing episodes of depression and (hypo)mania, which can also include psychotic symptoms such as delusions and hallucinations. References herein to BD are intended to include both bipolar disorder I (where there has been at least one manic episode) and bipolar disorder II (where there has been at least one hypomanic episode and one major depressive episode).

[0073] The term "diagnosis" as used herein encompasses identification, confirmation, and / or characterisation of BD, or predisposition thereto. The term "prognosis" as used herein encompasses the prediction of whether a patient it likely to develop BD. By "predisposition" it is meant that a subject does not currently present with the disorder, but is liable to be affected by the disorder in time, for example, develop the hypomanic or manic episodes typically characterised by BD.

[0074] The term "biomarker" means a distinctive biological or biologically derived indicator of a process, event, or condition. Biomarkers can be used in methods of diagnosis, e.g. clinical screening, and prognosis assessment and in monitoring the results of therapy, identifying patients most likely to respond to a particular therapeutic treatment, drug screening and development. Biomarkers and uses thereof are valuable for identification of new drug treatments and for discovery of new targets for drug treatment.

[0075] Data provided herein demonstrates that the biomarkers contain statistically significant and sensitive biomarkers for the diagnosis of BD, and for differential diagnosis of BD in an individual presenting with low mood. The individual may have previously been diagnosed with MDD.

[0076] The applicant previously carried out a study ("The Delta Study") conducted in the UK between April 27 , 2018, and February 6, 2020 to improve the diagnosis of mood disorders in individuals presenting with low mood. Further analysis of the results from The Delta Study confirmed high under-and misdiagnosis rates of mood disorders in individuals presenting with low mood, potentially leading to worsening of symptoms and decreased well-being, indicating the need for improved mental health triage in primary care (Martin-Key et al. (2021a)).

[0077] The applicant has since carried out additional work to identify a reproducible metabolomic biomarker signature in patient dried blood spots (DBSs) that enables diagnosis or prognosis of BD. In particular it differentiates BD from MDD during depressive episodes. The applicant has also found that the biomarkers add value when combined with self-reported patient information.

[0078] Samples and data from the Delta Study were utilized. The primary objective was to identify BD in patients with a recent (<5 years) diagnosis of MDD and current depressive symptoms (Patient Health Questionnaire-9 score>5). Participants were recruited online through voluntary response sampling. The analysis was carried out between February 2022 and June 2023.

[0079] Patient data were collected using a purpose-built online questionnaire (n=635 questions). DBS metabolites (n=630) were analysed using a targeted mass spectrometry-based platform (Biocrates MxP Quant 500). Mood disorder diagnoses were established using the Composite International Diagnostic Interview (CIDI). DBS metabolite levels were assessed in 241 depressed patients with a recent diagnosis of MDD, of whom 67 were subsequently diagnosed with BD by the CIDI (mean age, 28.0 [SD=7.8] years; 40 females [60%]) and 174 were confirmed as MDD (mean age, 28.1 [SD=6.8] years; 130 females [75%]). The identified 17-biomarker panel provided a cross-validated area under receiver operating characteristic curve (AUROC) of 0.71 (SD=0.12, P=7.3xl0'7), with ceramide (dl8:0 / 24: l) emerging as the strongest biomarker. Combining biomarker data with patient-reported information significantly enhanced diagnostic performance of models based on extensive demographic data, PHQ-9 scores, and the outcomes from the Mood Disorder Questionnaire. The identified biomarkers correlated primarily with lifetime manic symptoms, and validated in a separate group of patients who received a new clinical diagnosis of MDD (n=21) or BD (n=9) during the study's one-year follow-up period, with an AUROC of 0.73 (SD=0.06, P=4.9xl0’15).

[0080] The applicant's work provides a proof-of-concept for developing an accessible biomarker test to facilitate the differential diagnosis of BD and MDD, and highlights the potential involvement of ceramides in the pathophysiological mechanisms of mood disorders.

[0081] Figure 1 illustrates an embodiment of the invention. Prior to carrying out the method, a dried blood spot sample is obtained from a patient. The patient may be suffering from low mood, and / or may have previously been diagnosed with MDD. Conveniently the sample can be obtained by a pin-prick, which the patient can administer themself at home. The patient may be supplied with a lancet to obtain the blood sample. The patient will have been supplied with a dried blood spot collection card onto which the blood samples will be deposited and instructions for taking, preparing and returning the sample. Such cards are known in the art. Preferably the patient provides five dried blood spots to ensure that there is excess sample material. However, a single blood spot is sufficient. The dried blood spot sample 10 can then be mailed 12 to a laboratory for biomarker analysis.

[0082] In an embodiment, the amount of at least the following 17 biomarkers in the patient's sample is measured in a measuring step 14:

[0083] Cer(dl8:0 / 24: 1)

[0084] TG(18: 1_38:5) TrpBetaine

[0085] PC ae C36:5

[0086] 3-IAA

[0087] Lac

[0088] PC aa C32:3

[0089] GCA

[0090] Asn

[0091] TMAO 3-IPA t4-OH-Pro TG(17: 1_36:3) SM (OH) C14: l SM C20:2 PC aa 042:0 HexCer(d 18: 1 / 22:0)

[0092] These biomarkers are particularly useful as not only are they determinative of BD as opposed to MDD, they also can be conveniently measured from a dried blood spot from a patient. The measuring can be carried out by appropriate mass spectrometry techniques.

[0093] In a first stage, in a determining step 16, the probability of having BD is determined from each biomarker. The diagnostic model takes biomarker concentrations, measured on a continuous scale (rather than biomarker concentration thresholds), and converts them into corresponding probabilities of having BD. These probabilities are then summarized in a calculating step 18 across the biomarkers to calculate the overall probability of having BD. The final probability can range from 0 to 1. This final probability can then be compared in a comparing step 13 to the threshold determined when training the diagnostic model (for example, threshold probability = 0.2956886) to evaluate whether a person is more likely to have bipolar disorder (probability equal to or above the threshold) or major depressive disorder (probability below the threshold), allowing a diagnosis 15 of BD or MDD to be made. An advantage of being able to determine the probability of BD by comparing to a threshold value is that a control sample is not required. In some embodiments, for example, where the measurement platform does not provide absolute quantification without batch-to-batch variation, the amounts of the biomarkers may be compared in the comparing step 13 to those found in a control sample acting as a reference or quality control. This would preferably be any dried blood spot sample that remains stable over time, and is preferably prepared in large amounts, ideally from a healthy individual.

[0094] In other embodiments, the amounts of fewer than all 17 of the above biomarkers are measured. In some embodiments, at least eight of the biomarkers are measured. It is preferred that at least Cer(dl8:0 / 24: 1) is measured. In an embodiment, Cer(dl8:0 / 24: 1) is measured along with at least one of TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA, Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), PC aa C42:0, HexCer(dl8: 1 / 22:0). In an embodiment, Cer(dl8:0 / 24: l) is measured along with at least one of: TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA. In a particular embodiment, at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, and GCA are measured.

[0095] Figures 2 to 18 illustrate the relationship between biomarker concentrations (expressed as standardised concentration values on the horizontal axis) and the likelihood of a diagnosis of BD (SHapley Additive ExPlanations (SHAP) values on the vertical axis). SHAP values correspond to the log-odds of having a diagnosis of bipolar disorder, i.e. values above 0 indicate a higher likelihood of having bipolar disorder diagnosis, while values below 0 indicate a higher likelihood of having a diagnosis of major depressive disorder. Data points are coloured according to their SHAP value. Trend lines were fitted using LOESS (locally estimated scatterplot smoothing) regression. Biomarker concentration values (expressed in micromolar units [pM]) are given on the top horizontal axis. By way of example, the plots show that the cutoff for Cer(dl8:0 / 24: 1) would be approximately 0.9 pM.

[0096] The plots shown in Figures 2 to 18 demonstrate that Cer(dl8:0 / 24: l), PC ae C36:5, 3-IAA, Lac, GCA, Asn, TMAO, t4-OH-Pro and PC aa C42:0 are positively correlated with BD and that TG(18: 1_38:5), TrpBetaine, PC aa C32:3, 3-IPA, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2 and HexCer(dl8: 1 / 22:0) are negatively correlated with BD. The applicant has identified and validated a biomarker panel of 17 biomarkers, which demonstrated robust predictive performance. Importantly, the panel had a good predictive performance to differentiate between recent onset MDD patients and individuals with recent onset BD misdiagnosed as MDD, during depressive episodes.

[0097] Therefore, the biomarker panels described herein may be used as a panel of biomarkers for the differential diagnosis of BD from a further psychiatric disorder, such as MDD, in particular recent onset MDD. This aspect of the invention has the advantage of being able to diagnose whether patients that have previously been diagnosed with MDD, are likely to be misdiagnosed BD patients, and importantly, this diagnosis can take place during depressive episodes.

[0098] It will be appreciated that the term "differential diagnosis" refers to the positive diagnosis of BD from that of a further psychiatric disorder, such as MDD, in particular recent onset MDD.

[0099] It should be noted that references to biomarker amounts or levels also include references to a biomarker range. It will be appreciated that references herein to "difference in the amount" refer to either a higher or lower amount or level of the biomarker(s) in the test biological sample compared with a control, which may be one or more reference samples.

[0100] In an embodiment, the higher or lower level is a < 1-fold difference relative to the control or reference sample, such as a fold difference of 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1, 0.05, 0.01 or any ranges therebetween. In an embodiment, the lower level is between a 0.1 and 0.9 fold difference, such as between a 0.2 and 0.5 fold difference, relative to the control or reference sample.

[0101] In an embodiment, the higher or lower level is a > 1-fold difference relative to the control or reference sample, such as a fold difference of 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5, 9.0, 9.5, 10, 10.5, 11, 11.5, 12, 12.5, 15 or 20 or any ranges therebetween. In an embodiment, the higher level is between a 1 and 15 fold difference, such as between a 2 and 10 fold difference, relative to the control or reference sample. The control could be any suitable control, such as the level or amount of biomarker found in a normal control sample from a normal subject, a normal biomarker level; a normal biomarker range, the level in a sample from a subject with BD, or a diagnosed predisposition thereto; BD biomarker level, or BD biomarker range, or the level in a sample from a subject with MDD.

[0102] The individual may be a drug naive MDD patient (e.g. a first onset drug-naive patient). In an embodiment, the individual is first-onset or recent-onset drug naive MDD patient. In an embodiment, the individual is an unmedicated MDD patient. In an embodiment, the individual is an individual who has not yet experienced a hypomanic or manic episode. The method can be used to monitor a patient previously diagnosed as having BD, for example, to monitor the effectiveness of medication that the individual has been taking.

[0103] It will be understood that the term "drug naive" patients includes patients that have not previously been diagnosed or medicated for bipolar disorder. It will also be understood that the term "unmedicated" refers to patients that have not been taking medication for bipolar disorder (i.e. mood stabilizer or anti-psychotic medication) for at least 1 year, for example for at least 2, 3, 4, 5, 6, 7, 8, 9 or 10 years, in particular for at least 3 years.

[0104] Monitoring methods of the invention can be used to monitor onset, progression, stabilisation, amelioration and / or remission. In methods of diagnosing, prognosing or monitoring, detecting and / or quantifying the biomarker in a biological sample from a test subject may be performed on two or more occasions. Comparisons may be made between the level of biomarker in samples taken on two or more occasions.

[0105] Assessment of any change in the level of the biomarker in samples taken on two or more occasions may be performed. Modulation of the biomarker level is useful as an indicator of the state of bipolar disorder or predisposition thereto. An increase in the level of the biomarker over time is indicative of onset or progression, i.e. worsening of this disorder, whereas a decrease in the level of the biomarker indicates amelioration or remission of the disorder, or vice versa (depending on the biomarker in question).

[0106] Monitoring methods may also be used to monitor the efficacy of a therapy for BD in a subject having such a disorder, suspected of having such a disorder, or of being predisposed thereto. The method may further comprise comparing the level of the biomarker present in the test sample with one or more reference(s) and / or with one or more previous test sample(s) taken earlier from the same test subject, e.g. prior to commencement of therapy, and / or from the same test subject at an earlier stage of therapy. The method may comprise detecting a change in the level of the biomarker in test samples taken on different occasions.

[0107] For biomarkers that are increased in individuals with or having a predisposition to BD, a higher level or amount of the biomarker in the test sample is indicative of the presence of BD, or predisposition thereto; an equivalent or lower level or amount of the biomarker in the test sample is indicative of absence of BD and / or absence of a predisposition thereto. Examples of the biomarkers of the invention which have a higher level or amount in an individual with BD or a predisposition thereto include: 3-IAA, Asn, Cer(dl8:0 / 24: 1), GCA, Lac, PC aa C42:0, PC ae C36:5, t4-OH-Pro, and TMAO.

[0108] For biomarkers which are decreased in individuals with BD, a lower level of the biomarker in the test sample is indicative of the presence of BD, or predisposition thereto; an equivalent or higher level of the biomarker in the test sample is indicative of absence of BD and / or absence of a predisposition thereto. Examples of the biomarkers of the invention which have a lower level in an individual with BD or a predisposition thereto include: 3-IPA, HexCer(dl8: 1 / 22:0), PC aa C32:3, SM (OH) C14: l, SM C20:2, TG(17: 1_36:3), TG(18: 1_38 : 5) and TrpBetaine.

[0109] Efficient diagnosis, prognosis and monitoring methods provide very powerful "patient solutions" with the potential for improved prognosis, by establishing the correct diagnosis, allowing rapid identification of the most appropriate treatment (thus lessening unnecessary exposure to harmful drug side effects), reducing "down-time" and relapse rates.

[0110] Methods for monitoring efficacy of a therapy can be used to monitor the therapeutic effectiveness of existing therapies and new therapies in human subjects and in nonhuman animals (e.g. in animal models). These monitoring methods can be incorporated into screens for new drug substances and combinations of substances. Suitably, the time elapsed between taking samples from a subject undergoing diagnosis or monitoring will be 3 days, 5 days, a week, two weeks, a month, 2 months, 3 months, 6 or 12 months. Samples may be taken prior to and / or during and / or following therapy for bipolar disorder. Samples can be taken at intervals over the remaining life, or a part thereof, of a subject. The term "detecting" as used herein means confirming the presence of the biomarker present in the sample. Measuring the amount of the biomarker present in a sample may include determining the concentration of the biomarker present in the sample. Detecting and / or quantifying may be performed directly on the sample, or indirectly on an extract therefrom, or on a dilution thereof.

[0111] Detecting and / or quantifying can be performed by any method suitable to identify the presence and / or amount of a specific biomarker in a biological sample from a patient or a purification or extract of a biological sample or a dilution thereof. In methods of the invention, quantifying may be performed by measuring the concentration of the biomarker in the sample or samples. Biological samples that may be tested in a method of the invention include whole blood, blood serum, plasma, peripheral blood mononuclear cells, cerebrospinal fluid (CSF), urine, saliva, or other bodily fluid (tear fluid, synovial fluid, sputum), stool, breath, e.g. as condensed breath, or an extract or purification therefrom, or dilution thereof. Biological samples also include tissue homogenates, tissue sections and biopsy specimens from a live subject, or taken postmortem. The samples can be prepared, for example where appropriate diluted or concentrated, and stored in the usual manner.

[0112] In an embodiment, the biological sample is whole blood, blood serum or plasma, such as blood serum. It is particularly envisaged that the methods be carried out on a dried blood spot from an individual. This means that the patient can take their own sample remotely and then return it (for example, by mail) for analysis.

[0113] The biomarker may be detected using any suitable method known to the skilled person. It may be directly detected, for example using methods based on mass spectrometry. In some embodiments, the biomarker may be derivatised or converted into another molecule prior to detection by mass spectrometry or another suitable method. For example, detecting and / or quantifying can be performed by one or more method(s) selected from the group consisting of: SELDI (-TOF), MALDI (-TOF), a 1-D gel-based analysis such as capillary electrophoresis, mass spectrometry (MS) and tandem mass spectrometry (MS / MS) such as selected reaction monitoring (SRM), flow injection analysis (FIA), reverse phase (RP) LC, size permeation (gel filtration), ion exchange, affinity, HPLC, UPLC and other LC or LC MS-based techniques. Liquid chromatography (e.g. high pressure liquid chromatography (HPLC) or low pressure liquid chromatography (LPLC)), thin-layer chromatography, NMR (nuclear magnetic resonance) spectroscopy, Fourier transform infrared spectroscopy, or gas chromatography (GC) MS could also be used.

[0114] A kit for performing the method may include a questionnaire for use in diagnosing a patient with BD. The questionnaire may be used to support the results obtained by measuring the biomarkers or to help determine the severity of BD (i.e. severe, moderate or mild). The questionnaire may be the Hamilton Rating Scale for Depression (HAM-D, 17, 21 or 29 items) questionnaire. Other examples of suitable questionnaires which may be used, include: the Montgomery-Asberg Depression Rating Scale (MADRS), the Beck Depression Inventory (BDI), the Zung Self- Rating Depression Scale, the Wechsler Depression Rating Scale, the Raskin Depression Rating Scale, the Inventory of Depressive Symptomatology (IDS), the Quick Inventory of Depressive Symptomatology (QIDS), the Patient Health Questionnaire-9 (PHQ-9), the Patient Health Questionnaire-2 (PHQ-2), the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS), the Mood Disorder Questionnaire (MDQ), or other established or custom-built psychometric or demographic questionnaires.

[0115] Biomarker-based tests provide a first line assessment of 'new' patients, and provide objective measures for accurate and rapid diagnosis, in a time frame and with precision, not achievable using the current subjective measures.

[0116] Furthermore, diagnostic biomarker tests are useful to identify family members or patients at high risk of developing BD. This permits initiation of appropriate therapy, or preventive measures, e.g. managing risk factors. These approaches are recognised to improve outcome and may prevent overt onset of the disorder. Biomarker monitoring methods are also vital as patient monitoring tools, to enable the physician to determine whether relapse is due to worsening of the disorder, poor patient compliance or substance abuse. If pharmacological treatment is assessed to be inadequate, then therapy can be reinstated or increased; a change in therapy can be given if appropriate. If the biomarker is sensitive to the state of the disorder, it can provide an indication of the impact of drug therapy or of substance abuse.

[0117] In some embodiments, the levels of one or more biomarkers or the levels of a specific panel of biomarkers in a sample are compared to a reference standard ("reference standard" or "reference level") in order to direct treatment decisions. The reference standard used for any embodiment disclosed herein may comprise average, mean, or median levels of the one or more biomarkers or the levels of the specific panel of biomarkers in a control population. The reference standard may additionally comprise cutoff values or any other statistical attribute of the control population, such as a standard deviation from the mean levels of the one or more biomarkers or the levels of the specific panel of biomarkers.

[0118] In some embodiments, comparing the level of the one or more biomarkers is performed using a cutoff value. In related embodiments, if the level of the one or more biomarkers is greater than the cutoff value, the individual may be diagnosed as having, or being at risk of developing bipolar disorder. In other distinct embodiments, if the level of the one or more biomarkers is less than the cutoff value, the individual may be diagnosed as having, or being at risk of developing bipolar disorder. Cutoff values may be determined by statistical analysis of the control and disease populations to determine which levels represent a high likelihood that an individual does or does not belong to the control population. In some embodiments, comparing the amount of the one or more biomarkers is performed using other statistical methods. In related embodiments, comparing comprises logistic or linear regression, or machine learning methods such as Extreme Gradient Boosting (XGBoost). In other embodiments, comparing comprises computing an odds ratio.

[0119] In some embodiments, the control population may comprise healthy individuals, individuals with major depressive disorder, or individuals with bipolar disorder.

[0120] In some embodiments of the method, the method additionally includes the step of administering a BD medicament to a patient identified as having or of being predisposed to BD. Health practitioners treat bipolar depression by taking actions to ameliorate the causes or symptoms of the disorder in a patient. Treatment may comprise drug-based or non-drug-based therapies.

[0121] Drug-based therapies may include: selecting and administering one or more BD drugs to the patient, adjusting the dosage of a BD drug, adjusting the dosing schedule of a BD drug, and adjusting the length of the therapy with a BD drug. Appropriate drugs are selected by practitioners based on the nature of the symptoms and the patient's response to any previous treatments. The dosage of a BD drug can be adjusted as well by the practitioner based on the nature of the drug, the nature of the patient's symptoms, the patient's response to previous treatment, and the patient's response to the drug. The dosing schedule can also be adjusted by the practitioner based on the nature of the drug, the nature of the patient's symptoms, the patient's response to previous treatment, and the patient's response to the drug. Also, the length of the therapy can be adjusted by the practitioner based on the nature of the drug, the nature of the patient's symptoms, the patient's response to previous treatment, and the patient's response to the drug.

[0122] Additionally, the practitioner can select between a single drug therapy, a dual drug therapy, or a triple drug therapy. In some embodiments, a practitioner may optionally treat the patient with a combination of one or more drugs and one or more non-drug- based therapies. In an embodiment, the practitioner begins BD therapy based on a comparison between a reference level and the levels of one or more biomarkers or the levels of a specific panel of biomarkers in a sample from a patient.

[0123] The therapy may comprise the selection and administration of a drug to the patient by the practitioner. In another embodiment, therapy comprises the selection and administration of two drugs to the patient by the practitioner as part of dual therapy. In another embodiment, therapy comprises the selection and administration of three drugs to the patient by the practitioner as part of triple therapy.

[0124] Mood stabilizing drugs such as lithium are the most effective treatment for BD. The diagnosis of BD or a predisposition to BD based on the methods described herein enables a mood stabilizing medication to be prescribed. Mood stabilizing medication includes different drug classes such as lithium, anticonvulsant medication like valproate, carbamazepine or lamotrigine, and antipsychotic medicine such as haloperidol, olanzapine, quetiapine, aripiprazole or risperidone. Treatment with a combination of mood stabilizing medication and antidepressant medication is also sometimes indicated.

[0125] In some embodiments, treatment comprises administering to an individual a selective serotonin reuptake inhibitor ("SSRI"). In some embodiments, the SSRI is citalopram. In some embodiments, the SSRI is escitalopram. In some embodiments, the SSRI is fluoxetine. In some embodiments, the SSRI is paroxetine. In some embodiments, the SSRI is sertraline.

[0126] In other embodiments, treatment comprises administering to an individual a serotonin-norepinephrine reuptake inhibitors ("SNRI"). In some embodiments, the SNRI is venlafaxine. In other embodiments, the SNRI is duloxetine. In other embodiments, treatment comprises administering to an individual a norepinephrine and dopamine reuptake inhibitor ("NDRI"). In one embodiment, the NDRI is bupropion.

[0127] In other embodiments, treatment comprises administering to an individual a tetracyclic antidepressant ("tetracyclic"). In some embodiments, the tetracyclic is amoxapine. In some embodiments, the tetracyclic is maprotiline. In some embodiments, the tetracyclic is mazindol. In some embodiments, the tetracyclic is mirtazapine. In other embodiments, treatment comprises administering to an individual a tricyclic antidepressant ("tricyclic"). In some embodiments, the tricyclic is amitriptyline. In some embodiments, the tricyclic is imipramine. In some embodiments, the tricyclic is nortriptyline.

[0128] In other embodiments, treatment comprises administering to an individual a monoamine oxidase inhibitor ("MAOI"). In some embodiments, the MAOI is selegiline. In some embodiments, the MAOI is isocarboxazid. In some embodiments, the MAOI is phenelzine. In some embodiments, the MAOI is tranylcypromine.

[0129] In addition to or in lieu of drug-based therapies, in some embodiments a practitioner may also treat an individual with non-drug-based mood stabilizing therapies. In some embodiments, the non-drug based therapy comprises cognitive-behavioural therapy. In some embodiments, the non-drug based therapy comprises psychotherapy. In a related embodiment, the non-drug based therapy comprises psychodynamic therapy. In some embodiments, the non-drug based therapy comprises electroconvulsive therapy. In some embodiments, the non-drug based therapy comprises hospitalization and residential treatment programs. In some embodiments, the non- drug based therapy comprises vagus nerve stimulation. In some embodiments, the non-drug based therapy comprises transcranial magnetic stimulation. In some embodiments, the non-drug based therapy comprises regular, vigorous exercise. In one embodiment, the practitioner adjusts the mood stabilizing therapy based on a comparison between a reference level and the levels of one or more biomarkers or the levels of a specific panel of biomarkers in a sample from a patient. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different combination of drugs. In one embodiment, the practitioner adjusts the therapy by adjusting drug dosage. In one embodiment, the practitioner adjusts the therapy by adjusting dose schedule. In one embodiment, the practitioner adjusts the therapy by adjusting length of therapy. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug combination and adjusting drug dosage. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug combination and adjusting dose schedule. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug combination and adjusting length of therapy. In one embodiment, the practitioner adjusts the therapy by adjusting drug dosage and dose schedule. In one embodiment, the practitioner adjusts the therapy by adjusting drug dosage and adjusting length of therapy. In one embodiment, the practitioner adjusts the therapy by adjusting dose schedule and adjusting length of therapy. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug, adjusting drug dosage, and adjusting dose schedule. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug, adjusting drug dosage, and adjusting length of therapy. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug, adjusting dose schedule, and adjusting length of therapy. In one embodiment, the practitioner adjusts the therapy by adjusting drug dosage, adjusting dose schedule, and adjusting length of therapy. In one embodiment, the practitioner adjusts the therapy by selecting and administering a different drug, adjusting drug dosage, adjusting dose schedule, and adjusting length of therapy.

[0130] In some embodiments, treatment comprises a less aggressive therapy than a reference therapy. In one embodiment a less aggressive therapy comprises not administering drugs and taking a "watchful waiting" approach. In one embodiment a less aggressive therapy comprises delaying treatment. In one embodiment a less aggressive therapy comprises selecting and administering less potent drugs. In one embodiment a less aggressive therapy comprises decreasing dosage of mood stabilizing drugs. In one embodiment a less aggressive therapy comprises decreasing the frequency treatment. In one embodiment a less aggressive therapy comprises shortening length of therapy. In one embodiment, less aggressive therapy comprises selecting and administering less potent drugs and decreasing drug dosage. In one embodiment, less aggressive therapy comprises selecting and administering less potent drugs and decelerating dose schedule. In one embodiment, less aggressive therapy comprises selecting and administering less potent drugs and shortening length of therapy. In one embodiment, less aggressive therapy comprises decreasing drug dosage and decelerating dose schedule. In one embodiment, less aggressive therapy comprises decreasing drug dosage and shortening length of therapy. In one embodiment, less aggressive therapy comprises decelerating dose schedule and shortening length of therapy. In one embodiment, less aggressive therapy comprises selecting and administering less potent drugs, decreasing drug dosage, and decelerating dose schedule. In one embodiment, less aggressive therapy comprises selecting and administering less potent drugs, decreasing drug dosage, and shortening length of therapy. In one embodiment, less aggressive therapy comprises selecting and administering less potent drugs, decelerating dose schedule, and shortening length of therapy. In one embodiment, less aggressive therapy comprises decreasing drug dosage, decelerating dose schedule, and shortening length of therapy. In one embodiment, less aggressive therapy comprises selecting and administering less potent drugs, decreasing drug dosage, decelerating dose schedule, and shortening length of therapy. In some embodiments, a less aggressive therapy comprises administering only non-drug-based therapies.

[0131] Treatment may comprise a more aggressive therapy than a reference therapy. In one embodiment a more aggressive therapy comprises earlier administration of mood stabilizing drugs. In one embodiment a more aggressive therapy comprises increased dosage of mood stabilizing drugs. In one embodiment a more aggressive therapy comprises increased length of therapy. In one embodiment a more aggressive therapy comprises increased frequency of the dose schedule. In one embodiment, more aggressive therapy comprises selecting and administering more potent drugs and increasing drug dosage. In one embodiment, more aggressive therapy comprises selecting and administering more potent drugs and accelerating dose schedule. In one embodiment, more aggressive therapy comprises selecting and administering more potent drugs and increasing length of therapy. In one embodiment, more aggressive therapy comprises increasing drug dosage and accelerating dose schedule. In one embodiment, more aggressive therapy comprises increasing drug dosage and increasing length of therapy. In one embodiment, more aggressive therapy comprises accelerating dose schedule and increasing length of therapy. In one embodiment, more aggressive therapy comprises selecting and administering more potent drugs, increasing drug dosage, and accelerating dose schedule. In one embodiment, more aggressive therapy comprises selecting and administering more potent drugs, increasing drug dosage, and increasing length of therapy. In one embodiment, more aggressive therapy comprises selecting and administering more potent drugs, accelerating dose schedule, and increasing length of therapy. In one embodiment, more aggressive therapy comprises increasing drug dosage, accelerating dose schedule, and increasing length of therapy. In one embodiment, more aggressive therapy comprises selecting and administering more potent drugs, increasing drug dosage, accelerating dose schedule, and increasing length of therapy. In some embodiments, a more aggressive therapy comprises administering a combination of drug-based and non-drug-based therapies.

[0132] The results of the analyses will often be communicated to physicians and / or patients (or other interested parties such as researchers) in a transmittable form that can be communicated or transmitted to any of the above parties. Such a form can vary and can be tangible or intangible. The results can be embodied in descriptive statements, diagrams, photographs, charts, images or any other visual forms. The statements and visual forms can be recorded on a tangible medium such as papers, computer readable media such as hard disks, compact disks, USB, SD card etc., or on an intangible medium, e.g., an electronic medium in the form of email or website on internet or intranet. In addition, results can also be recorded in a sound form and transmitted through any suitable medium, e.g., analogue or digital cable lines, fibre optic cables, etc., via telephone, facsimile, wireless mobile phone, internet phone and the like.

[0133] Thus, the information and data on a test result can be produced anywhere in the world and transmitted to a different location. As an illustrative example, when an assay is conducted outside a given territory, the information and data on a test result may be generated, cast in a transmittable form as described above, and then imported into the given territory.

[0134] Accordingly, embodiments of the present invention also encompass a method for producing a transmittable form of information on levels of one or more biomarkers or levels of a specific panel of biomarkers for at least one patient sample. The method comprises the steps of (1) determining levels of one or more biomarkers or levels of a specific panel of biomarkers for at least one patient sample according to methods of the present invention; and (2) embodying the result of the determining step in a transmittable form. The transmittable form is the product of such a method.

[0135] Techniques for analysing levels of one or more biomarkers or levels of a specific panel of biomarkers for at least one patient sample will often be implemented using hardware, software or a combination thereof in one or more computer systems or other processing systems capable of effectuating such analysis.

[0136] Thus, embodiments of the present invention further provide a system for determining whether an individual suffers from bipolar disorder, comprising: (1) a sample analyser for determining the levels of one or more biomarkers or levels of a specific panel of biomarkers for at least one patient sample, wherein the sample analyser contains the patient sample; (2) a first computer program for (a) receiving data regarding the levels of one or more biomarkers or the levels of a specific panel of biomarkers; and optionally (3) a second computer program for comparing the test value to one or more reference standards each associated with a predetermined degree of risk of bipolar disorder.

[0137] The analysis function can also be embodied in computer program products and used in the systems described above or other computer- or internet-based systems. Accordingly, another aspect of the present invention relates to a computer program product comprising a computer-usable medium having computer-readable program codes or instructions embodied thereon for enabling a processor to carry out disease risk analysis. These computer program instructions may be loaded onto a computer or other programmable apparatus to produce a machine, such that the instructions, which execute on the computer or other programmable apparatus, create means for implementing the functions or steps described above. These computer program instructions may also be stored in a computer-readable memory or medium that can direct a computer or other programmable apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory or medium produce an article of manufacture including instructions which implement the analysis. The computer program instructions may also be loaded onto a computer or other programmable apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions or steps described above.

[0138] Thus one aspect of the present invention provides a system for determining whether a patient has BD. Generally speaking, the system comprises (1) computer program for receiving, storing, and / or retrieving data regarding levels of biomarkers in a patient's sample and optionally clinical parameter data (e.g., disease-related symptoms); (2) computer program for querying this patient data; (3) computer program for concluding whether an individual suffers from bipolar disorder based on this patient data; and optionally (4) computer program for outputting / displaying this conclusion. In some embodiments this computer program for outputting the conclusion may comprise a computer program for informing a health care professional of the conclusion.

[0139] All optional and preferred features and modifications of the described embodiments and dependent claims are usable in all aspects of the invention taught herein. Furthermore, the individual features of the dependent claims, as well as all optional and preferred features and modifications of the described embodiments are combinable and interchangeable with one another. EXAMPLES

[0140] Example 1 - Differential metabolomic signature of bipolar and unipolar depression

[0141] Methods

[0142] Study participants

[0143] Analysed samples and data were collected as part of the Delta Study carried out by the Cambridge Centre for Neuropsychiatric Research (CCNR) at the University of Cambridge, UK (International Registered Report Identifier: DERR1-10.2196 / 18453) (Olmert et al. (2020)). The Delta Study aimed to improve diagnosis of mood disorders in patients presenting with depressive symptoms using blood biomarker and digital questionnaire data (Han et al. (2020); Olmert et al. (2020); Benacek et al. (2021); Martin-Key et al. (2021a); Martin-Key et al. (2021b); Mirea et al. (2021); Tomasik et al. (2021); Benacek et al. (2022)). The primary objective of the Delta Study was to identify bipolar disorder (BD) in patients who had been diagnosed with major depressive disorder (MDD) within the previous 5 years. Participants were recruited through email, the CCNR website, and paid Facebook advertisements. Eligibility criteria included age between 18 and 45 years old, residency of the United Kingdom, at least mild current depressive symptoms (Patient Health Questionnaire-9 [PHQ-9] (Kroenke et al. (2001)) total score >5), not pregnant or breastfeeding, and not suicidal. Recruitment started on April 27 , 2018, and the study closed on February 6, 2020. All participants provided written, digitally signed informed consent. The study was approved by the University of Cambridge Human Biology Research Ethics Committee (approval number HBREC 2017.11) and complied with ethical standards of human experimentation as described in the Declaration of Helsinki (World Medical Association (2013), the Good Clinical Practice guidelines, and ISO 14155:2011.

[0144] Procedures

[0145] Enrolled participants were asked to complete a purpose-built online mental health questionnaire. The questionnaire was developed in collaboration with experienced psychiatrists and a service user advisory group, and was based on existing structured diagnostic interviews as well as a range of mental health screening questionnaires, such as the Mood Disorder Questionnaire (MDQ) (Hirschfeld et al. (2000)) and the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS) (Tennant et al. (2007)). The questionnaire consisted of 635 questions split into 6 modules: 1) demographic information, 2) manic and hypomanic symptoms, 3) depressive symptoms, 4) personality traits (based on the Big Five inventory (Goldberg (1993))), 5) psychiatric history, and 6) comorbid psychiatric symptoms. The questionnaire was adaptive to answers given by participants, so that only relevant questions were asked, and the maximum possible number of questions asked to an individual was 382.

[0146] Following completion of the online questionnaire, eligible participants who consented to providing a blood sample and completing a telephone diagnostic interview, who were free from blood-borne illnesses, and who had no previous diagnosis of schizophrenia, were provided with a dried blood spot (DBS) collection kit by post. The kit was designed to allow minimally invasive blood sample collection in a non-clinical setting, and was a Conformite Europeenne (CE)-marked device under Article 22 of the Medical Device Regulation 2017 / 745. The kit included pre-injection cleaning swabs, sterile finger prick lancets, a DBS collection card (226 Spot Saver Cards, PerkinElmer), adhesive plasters, and cotton pads. Detailed instructions for DBS sample collection were provided in a leaflet and as an online video. Participants were asked to spot 5 separate DBSs onto the card, after at least 6 hours of fasting, and allow the card to dry for a minimum of 3 hours at room temperature. Cards were then placed in the provided resealable bags with desiccant, returned by post using pre-paid envelopes, and stored at room temperature in a humidity-controlled cabinet until analysis.

[0147] Outcomes

[0148] Participants who successfully completed the online questionnaire and returned the DBS sample were invited to complete the World Health Organization World Mental Health Composite International Diagnostic Interview (CIDI), version 3.0 (Kessler & Ustun (2004)), via telephone. The CIDI is a modular diagnostic tool widely used in epidemiological studies on mental health (Kessler et al. (2005)) that shows good concordance with structured diagnostic interviews conducted by clinicians (Haro et al. (2006)). All interviewers conducting the CIDI received in-person training from an external CIDI-certified instructor, and internal training and mentoring. Only modules of the CIDI required for the lifetime diagnosis of mood disorder, i.e. the screening, depression, and mania sections, were implemented. We adopted voluntary response sampling, with the CIDI interviews continuing until the pre-specified study recruitment targets were met (Olmert et al. (2020)). Follow-up data regarding changes in diagnosis, treatment, and wellbeing after enrollment were collected through an online questionnaire at 6 and 12 months.

[0149] A subset of samples was selected for the present analysis based on the outcomes of the CIDI and changes in diagnosis at 6 or 12 months. The discovery cohort included participants diagnosed with MDD within the previous five years, whose diagnosis was either confirmed as MDD or altered to BD using the CIDI. The inventors focused on the BD I subtype of BD, to increase the chances of detecting a robust biomarker signal. The validation cohort comprised of patients initially diagnosed as having MDD, excluded from the discovery cohort, who received a diagnosis of BD from a healthcare professional during the 1-year follow-up, as well as depressed individuals without a mood disorder diagnosis at baseline who were diagnosed with MDD by a healthcare professional during the follow-up period. Participants whose DBS samples were not usable, or who provided inconsistent answers to the screening question about elevated mood on the online questionnaire and the telephone interview, were excluded from analysis to ensure data integrity.

[0150] Absolute concentrations of 630 metabolites from 22 biochemical classes were measured in the DBS samples using a targeted mass spectrometry (MS)-based metabolomic platform (MxP Quant 500, Biocrates Life Sciences AG). All analytical procedures were carried out in a quality-controlled and ISO 9001 :2015-certified laboratory (Biocrates Life Sciences AG, Innsbruck, Austria) as described in detail in WO 2007 / 003343 and WO 2007 / 003344, the contents of which are incorporated herein by reference. In brief, 3 mm DBS discs cut from the original DBS collection cards were transferred onto a 96-well filter plate impregnated with isotope-labelled internal standards. Next, phenyl isothiocyanate (PITC) solution was used to derivatize some of the analytes (e.g. amino acids), followed by extraction of compounds with an organic solvent, and a dilution step. The resulting extracts were analyzed by flow injection analysis-tandem mass spectrometry (FIA-MS / MS; lipids and hexoses) using a 5500 QTRAP instrument (AB Sciex, Darmstadt, Germany) equipped with an electrospray ionization (ESI) source, and by liquid chromatography-tandem mass spectrometry (LC-MS / MS; small molecules) using a Xevo TQ-XS instrument (Waters, Milford, MA, USA), in a multiple reaction monitoring (MRM) mode. Samples were randomized across plates and plate positions to minimize technical bias, and replicates and quality control samples were included to monitor variation in sample preparation and instrument performance. Experimenters were blind to sample diagnostic allocation.

[0151] Statistical analysis

[0152] Raw metabolite data were processed using the Sciex Analyst (AB Sciex, Darmstadt, Germany) and Waters MassLynx (Waters, Milford, MA, USA) software. Absolute quantification was performed based on isotope-labelled internal standards spiked into the samples and external calibration curves. Further data processing and analysis were conducted in R version 4.2.2 (https : / / w .r-proiect.org). The lower limit of quantitation (LLQ) was calculated as the maximum LLQ measured across experimental plates, and harmonized across the batches. Analytes with more than 50% of values missing in any of the batches, a high batch-to-batch variability in the proportion of missing values (>0.25), or affected by the fasting status (P<0.05; Mann-Whitney test with Bonferroni correction) were excluded from analysis. For the remaining analytes, potential batch effects were removed by robust scaling, i.e. subtracting batch median from individual values and dividing by batch interquartile range. Missing values were imputed with values 10% lower than the lowest scaled value. Principal component analysis (PCA) was used to identify outliers and confounding effects.

[0153] The online mental health questionnaire data were restructured so that answers to equivalent questions were concatenated (e.g., current and past symptoms), missing values were imputed where feasible (e.g., the number of relatives with depression was set to 0 for participants with no family history of mental health conditions), and features derived from the original variables were added (guided by the design of existing diagnostic algorithms, e.g. the number of concurrent symptoms or MDQ scores) (Tomasik et al. (2021)). Ordinal questionnaire data were converted to ranks, and categorical data were encoded as dummy variables. The online questionnaire features that were duplicated, bijections, or constant were excluded from analysis. Diagnostic models were built using the decision tree-based machine learning method, Extreme Gradient Boosting (XGBoost) (Chen & Guestrin (2016)), selected due to its capabilities in handling missing values, detecting non-linear relationships and interactions between variables, being robust to correlated features, and ensuring model explainability and interpretability. The models were trained based on 10-fold stratified cross-validation repeated ten times, with the number of folds and repeats selected for best replicability (Bouckaert & Frank (2004)). Tuned model parameters included the number of trees (1 to 100), tree depth (1 or 2, to allow for first order interactions), and the learning rate (0.1 or 0.3). Model performance was evaluated using the cross-validated area under the receiver operating characteristic curve (AUROC). Reported values are averages (± standard deviation [SD]) obtained from applying the cross-validated models. Feature importance was assessed by measuring gain, i.e. the increase in accuracy brought by a feature to the branches it appeared on, and cumulative importance was calculated as the sum of individual feature importances within feature categories. The directionality of biomarker associations was determined using the SHapley Additive exPlanations (SHAP) method (Lundberg & Lee (2017)). Decision curve analysis (Vickers & Elkin (2006)) was used to estimate the clinical utility (standardized net benefit) of predictive models with and without biomarkers at varying diagnostic thresholds. The optimal classification cut-off was determined using Youden's J statistic (Youden (1950)), to account for the imbalance between the groups. In the null model, which did not include any features, the cut-off was determined as the baseline prevalence. Model metrics were compared using the Nadeau and Bengio's corrected resampled t-test (Nadeau & Bengio (2003)) which adjusts for non-independence of the resampled statistics. The association between DBS biomarker levels and symptom data was measured using the absolute Pearson correlation coefficient, averaged within symptom categories. Clustered heatmaps were created based on maximum distances and complete linkage clustering. Figures were prepared in the R package ggplot2 and Inkscape v.1.2.2. Results

[0154] Study Overview The discovery cohort consisted of 241 patients with a recent diagnosis of MDD and current depressive symptoms, of whom 67 were subseguently diagnosed with BD by the CIDI and 174 were confirmed as MDD (see Table 1).

[0155] Table 1 - Demographic and clinical characteristics of study participants in the original cohort

[0156] The validation cohort included 30 participants who received a new diagnosis of MDD (n=21) or BD (n=9) during the study's one-year follow-up period (Table 2). Table 2 - Demographic and clinical characteristics of study participants in the validation cohort

[0157] The analysed data comprised 290 DBS metabolite readouts representing 19 biochemical classes (see Table 3), and 992 digital questionnaire features divided into 9 categories: demographics (n= lll features), psychiatric history (n=264), manic symptoms (n=142), depressive symptoms (n = 199), comorbid psychiatric symptoms (n=154), personality traits (n = 119), and single outcome items from the MDQ, PHQ-9, and WEMWBS.

[0158] Table 3 - Metabolites included in the analysis following data processing and quality control Alanine Ala Amino acids

[0159] Asparagine Asn Amino acids

[0160] Aspartic Acid Asp Amino acids

[0161] Cysteine Cys Amino acids

[0162] Glutamic Acid Glu Amino acids

[0163] Glycine Gly Amino acids

[0164] Isoleucine lie Amino acids

[0165] Leucine Leu Amino acids

[0166] Phenylalanine Phe Amino acids

[0167] Proline Pro Amino acids

[0168] Serine Ser Amino acids

[0169] Threonine Thr Amino acids

[0170] Tyrosine Tyr Amino acids

[0171] Valine Vai Amino acids

[0172] Deoxycholic acid DCA Bile acids

[0173] Glycocholic acid GCA Bile acids

[0174] Glycolithocholic acid sulfate GLCAS Bile acids

[0175] Taurocholic acid TCA Bile acids beta-Alanine beta-Ala Biogenic amines gamma-Amino-butyric acid GABA Biogenic amines

[0176] Hexose Hl Carbohydrates and related

[0177] Hippuric acid HipAcid Carboxylic acids

[0178] Lactic acid Lac Carboxylic acids

[0179] Ceramide (dl6: 1 / 18:0) Cer(dl6: Ceramides

[0180] Ceramide (dl6: 1 / 22:0) Cer(d l6: Ceramides

[0181] Ceramide (dl6: 1 / 24:0) Cer(d l6: Ceramides

[0182] Ceramide (dl8: 1 / 16:0) Cer(dl8: Ceramides

[0183] Ceramide (dl8: 1 / 18:0) Cer(dl8: Ceramides

[0184] Ceramide (dl8: 1 / 20:0(OH)) Cer(d l8: Ceramides

[0185] Ceramide (dl8: 1 / 20:0) Cer(d l8: Ceramides

[0186] Ceramide (dl8: 1 / 22:0) Cer(d l8: Ceramides

[0187] Ceramide (dl8: 1 / 23:0) Cer(d l8: Ceramides

[0188] Ceramide (dl8: 1 / 24:0) Cer(d l8: Ceramides

[0189] Ceramide (dl8: 1 / 24: 1) Cer(d l8: Ceramides

[0190] Ceramide (dl8: 1 / 26:0) Cer(d l8: Ceramides

[0191] Ceramide (d l8: 1 / 26: 1) Cer(d l8: Ceramides

[0192] Ceramide (d l8:2 / 16:0) Cer(d l8: Ceramides

[0193] Ceramide (d l8:2 / 18:0) Cer(d l8: Ceramides

[0194] Ceramide (d l8:2 / 20:0) Cer(d l8: Ceramides

[0195] Ceramide (d l8:2 / 22:0) Cer(d l8: Ceramides

[0196] Ceramide (d l8:2 / 24:0) Cer(d l8: Ceramides

[0197] Ceramide (dl8:2 / 24: 1) Cer(d l8: Ceramides

[0198] Dihexosylceramide (d l8: 1 / 14:0) Hex2Cer Ceramides

[0199] Dihexosylceramide (d l8: 1 / 16:0) Hex2Cer Ceramides Dihexosylceramide (dlS: 1 / 18:0) Hex2Cer(dl8: 1 / 18:0) Ceramides Dihexosylceramide (dl8: 1 / 20:0) Hex2Cer(dl8: 1 / 20:0) Ceramides Dihexosylceramide (dl8: 1 / 22:0) Hex2Cer(dl8: 1 / 22:0) Ceramides Dihexosylceramide (dl8: 1 / 24:0) Hex2Cer(dl8: 1 / 24:0) Ceramides Dihexosylceramide (dl8: 1 / 24: 1) Hex2Cer(dl8: 1 / 24: 1) Ceramides Dihexosylceramide (dl8: 1 / 26:0) Hex2Cer(dl8: 1 / 26:0) Ceramides Dihydroceramide (dl8:0 / 22:0) Cer(dl8:0 / 22:0) Ceramides Dihydroceramide (dl8:0 / 24:0) Cer(dl8:0 / 24:0) Ceramides Dihydroceramide (dl8:0 / 24: 1) Cer(dl8:0 / 24: 1) Ceramides Hexosylceramide (dl6: 1 / 22:0) HexCer(dl6: 1 / 22:0) Ceramides Hexosylceramide (dl6: 1 / 24:0) HexCer(dl6: 1 / 24:0) Ceramides Hexosylceramide (dl8: 1 / 14:0) HexCer(dl8: 1 / 14:0) Ceramides Hexosylceramide (dl8: 1 / 16:0) HexCer(dl8: 1 / 16:0) Ceramides Hexosylceramide (dl8: 1 / 18:0) HexCer(dl8: 1 / 18:0) Ceramides Hexosylceramide (dl8: 1 / 20:0) HexCer(dl8: 1 / 20:0) Ceramides Hexosylceramide (dl8: 1 / 22:0) HexCer(dl8: 1 / 22:0) Ceramides Hexosylceramide (dl8: 1 / 23:0) HexCer(dl8: 1 / 23:0) Ceramides Hexosylceramide (dl8: 1 / 24:0) HexCer(dl8: 1 / 24:0) Ceramides Hexosylceramide (dl8: 1 / 24: 1) HexCer(dl8: 1 / 24: 1) Ceramides Hexosylceramide (dl8:2 / 16:0) HexCer(dl8:2 / 16:0) Ceramides Hexosylceramide (dl8:2 / 18:0) HexCer(dl8:2 / 18:0) Ceramides Hexosylceramide (dl8:2 / 22:0) HexCer(dl8:2 / 22:0) Ceramides Trihexosylceramide (dl8: 1 / 16:0) Hex3Cer(dl8: 1 / 16:0) Ceramides Trihexosylceramide (dl8: 1 / 20:0) Hex3Cer(dl8: 1 / 20:0) Ceramides Trihexosylceramide (dl8: 1 / 22:0) Hex3Cer(dl8: 1 / 22:0) Ceramides Trihexosylceramide (dl8: 1 / 24: 1) Hex3Cer(dl8: 1 / 24: 1) Ceramides Trihexosylceramide (dl8: 1 / 26: 1) Hex3Cer(dl8: 1 / 26: 1) Ceramides Cholesteryl ester 14:0 CE(14:0) Cholesteryl esters Cholesteryl ester 14: 1 CE(14: 1) Cholesteryl esters Cholesteryl ester 15:0 CE(15:0) Cholesteryl esters Cholesteryl ester 16:0 CE(16:0) Cholesteryl esters Cholesteryl ester 17:0 CE(17:0) Cholesteryl esters Cholesteryl ester 17: 1 CE(17: 1) Cholesteryl esters Cholesteryl ester 18:0 CE(18:0) Cholesteryl esters Cholesteryl ester 18: 1 CE(18: 1) Cholesteryl esters Cholesteryl ester 18:2 CE(18:2) Cholesteryl esters Cholesteryl ester 18:3 CE(18:3) Cholesteryl esters Cholesteryl ester 20:0 CE(20:0) Cholesteryl esters Cholesteryl ester 20: 1 CE(20: l) Cholesteryl esters Cholesteryl ester 22:5 CE(22:5) Cholesteryl esters Cholesteryl ester 22:6 CE(22:6) Cholesteryl esters p-Cresol sulfate p-Cresol-SO4 Cresols Arachidonic acid AA Fatty acids

[0200] Docosahexaenoic acid DHA Fatty acids Eicosapentaenoic acid EPA Fatty acids Octadecadienoate FA(18:2) Fatty acids Octadecenoic acid FA(18: 1) Fatty acids Indoleacetic acid 3-IAA Indoles and derivatives Indolepropionic acid 3-IPA Indoles and derivatives Indoxyl sulfate Ind-SO4 Indoles and derivatives Lysophosphatidyl-choline a C14:0 lysoPC a C14:0 Lysophosphatidylcholines Lysophosphatidyl-choline a C16:0 lysoPC a C16:0 Lysophosphatidylcholines Lysophosphatidyl-choline a C16: 1 lysoPC a C16: 1 Lysophosphatidylcholines Lysophosphatidyl-choline a C17:0 lysoPC a C17:0 Lysophosphatidylcholines Lysophosphatidyl-choline a C18:0 lysoPC a C18:0 Lysophosphatidylcholines Lysophosphatidyl-choline a CIS: 1 lysoPC a CIS: 1 Lysophosphatidylcholines Lysophosphatidyl-choline a C18:2 lysoPC a C18:2 Lysophosphatidylcholines Lysophosphatidyl-choline a C20:3 lysoPC a C20:3 Lysophosphatidylcholines Lysophosphatidyl-choline a C20:4 lysoPC a C20:4 Lysophosphatidylcholines Lysophosphatidyl-choline a C24:0 lysoPC a C24:0 Lysophosphatidylcholines Lysophosphatidyl-choline a C26:0 lysoPC a C26:0 Lysophosphatidylcholines Lysophosphatidyl-choline a C26: l lysoPC a C26: l Lysophosphatidylcholines Lysophosphatidyl-choline a C28:0 lysoPC a C28:0 Lysophosphatidylcholines Lysophosphatidyl-choline a C28: 1 lysoPC a C28: 1 Lysophosphatidylcholines Hypoxanthine Hypoxanthine Nucleobases and related

[0201] Phosphatidyl-choline aa C24:0 PC aa C24:0 Phosphatidylcholines Phosphatidyl-choline aa C28: 1 PC aa C28: 1 Phosphatidylcholines Phosphatidyl-choline aa C30:0 PC aa C30:0 Phosphatidylcholines Phosphatidyl-choline aa C32:0 PC aa C32:0 Phosphatidylcholines Phosphatidyl-choline aa C32: 1 PC aa C32: 1 Phosphatidylcholines Phosphatidyl-choline aa C32:2 PC aa C32:2 Phosphatidylcholines Phosphatidyl-choline aa C32:3 PC aa C32:3 Phosphatidylcholines Phosphatidyl-choline aa C34: l PC aa C34: l Phosphatidylcholines Phosphatidyl-choline aa C34:2 PC aa C34:2 Phosphatidylcholines Phosphatidyl-choline aa C34:3 PC aa C34:3 Phosphatidylcholines Phosphatidyl-choline aa C34:4 PC aa C34:4 Phosphatidylcholines Phosphatidyl-choline aa C36:0 PC aa C36:0 Phosphatidylcholines Phosphatidyl-choline aa C36: 1 PC aa C36: 1 Phosphatidylcholines Phosphatidyl-choline aa C36:2 PC aa C36:2 Phosphatidylcholines Phosphatidyl-choline aa C36:3 PC aa C36:3 Phosphatidylcholines Phosphatidyl-choline aa C36:4 PC aa C36:4 Phosphatidylcholines

[0202] Phosphatidyl-choline aa C36: 5 PC aa C36: 5 Phosphatidylcholines Phosphatidyl-choline aa C36:6 PC aa C36:6 Phosphatidylcholines Phosphatidyl-choline aa C38:0 PC aa C38:0 Phosphatidylcholines Phosphatidyl-choline aa C38:3 Phosphatidylcholines Phosphatidyl-choline aa C38:4 Phosphatidylcholines Phosphatidyl-choline aa C38: 5 Phosphatidylcholines Phosphatidyl-choline aa C38:6 Phosphatidylcholines Phosphatidyl-choline aa C40:2 Phosphatidylcholines Phosphatidyl-choline aa C40:3 Phosphatidylcholines Phosphatidyl-choline aa C40:4 Phosphatidylcholines Phosphatidyl-choline aa C40: 5 Phosphatidylcholines Phosphatidyl-choline aa C40:6 Phosphatidylcholines Phosphatidyl-choline aa C42:0 Phosphatidylcholines Phosphatidyl-choline aa C42: l Phosphatidylcholines Phosphatidyl-choline aa C42:2 Phosphatidylcholines Phosphatidyl-choline aa C42:4 Phosphatidylcholines Phosphatidyl-choline aa C42: 5 Phosphatidylcholines Phosphatidyl-choline aa C42:6 Phosphatidylcholines Phosphatidyl-choline ae C30:0 Phosphatidylcholines Phosphatidyl-choline ae C30: 1 Phosphatidylcholines Phosphatidyl-choline ae C30:2 Phosphatidylcholines Phosphatidyl-choline ae C32: 1 Phosphatidylcholines Phosphatidyl-choline ae C32:2 Phosphatidylcholines Phosphatidyl-choline ae C34:0 Phosphatidylcholines Phosphatidyl-choline ae C34: l Phosphatidylcholines Phosphatidyl-choline ae C34:2 Phosphatidylcholines Phosphatidyl-choline ae C34:3 Phosphatidylcholines Phosphatidyl-choline ae C36:0 Phosphatidylcholines Phosphatidyl-choline ae C36: 1 Phosphatidylcholines Phosphatidyl-choline ae C36:2 Phosphatidylcholines Phosphatidyl-choline ae C36:3 Phosphatidylcholines Phosphatidyl-choline ae C36:4 Phosphatidylcholines Phosphatidyl-choline ae C36: 5 Phosphatidylcholines Phosphatidyl-choline ae C38:0 Phosphatidylcholines Phosphatidyl-choline ae C38: 1 Phosphatidylcholines Phosphatidyl-choline ae C38:2 Phosphatidylcholines Phosphatidyl-choline ae C38:3 Phosphatidylcholines Phosphatidyl-choline ae C38:4 Phosphatidylcholines Phosphatidyl-choline ae C38: 5 Phosphatidylcholines Phosphatidyl-choline ae C38:6 Phosphatidylcholines Phosphatidyl-choline ae C40: l Phosphatidylcholines Phosphatidyl-choline ae C40:2 Phosphatidylcholines Phosphatidyl-choline ae C40:3 Phosphatidylcholines Phosphatidyl-choline ae C40:4 Phosphatidylcholines Phosphatidyl-choline ae C40: 5 Phosphatidylcholines Phosphatidyl-choline ae C40:6 Phosphatidylcholines Phosphatidyl-choline ae C42: l Phosphatidylcholines Phosphatidyl-choline ae C42:2 Phosphatidylcholines Phosphatidyl-choline ae C42:3 Phosphatidylcholines Phosphatidyl-choline ae C42:4 Phosphatidylcholines Phosphatidyl-choline ae C44:3 Phosphatidylcholines Phosphatidyl-choline ae C44:4 Phosphatidylcholines Phosphatidyl-choline ae C44:5 Phosphatidylcholines Phosphatidyl-choline ae C44:6 Phosphatidylcholines Hydroxysphingo-myelin C14: 1 Sphingomyelins Hydroxysphingo-myelin C16: 1 Sphingomyelins Hydroxysphingo-myelin C22: l Sphingomyelins Hydroxysphingo-myelin C22:2 Sphingomyelins Hydroxysphingo-myelin C24: l Sphingomyelins Sphingomyelin C16:0 Sphingomyelins Sphingomyelin C 16 : 1 Sphingomyelins Sphingomyelin C18:0 Sphingomyelins Sphingomyelin C18: 1 Sphingomyelins Sphingomyelin C20:2 Sphingomyelins Sphingomyelin C24:0 Sphingomyelins Sphingomyelin C24: l Sphingomyelins Sphingomyelin C26:0 Sphingomyelins Sphingomyelin C26: l Sphingomyelins Triacylglyceride (14:0_34: 1) Triglycerides Triacylglyceride (14:0_36:2) Triglycerides Triacylglyceride (14:0_36:3) Triglycerides Triacylglyceride (16:0_30:2) Triglycerides Triacylglyceride (16:0_32: 1) Triglycerides Triacylglyceride (16:0_32:2) Triglycerides Triacylglyceride (16:0_33: 1) Triglycerides Triacylglyceride (16:0_34: 1) Triglycerides Triacylglyceride (16:0_34:2) Triglycerides Triacylglyceride (16:0_34:3) Triglycerides Triacylglyceride (16:0_35: 1) Triglycerides Triacylglyceride (16:0_35:2) Triglycerides Triacylglyceride (16:0_35:3) Triglycerides Triacylglyceride (16:0_36:2) Triglycerides Triacylglyceride (16:0_36:3) Triglycerides Triacylglyceride (16:0_36:4) Triglycerides Triacylglyceride (16:0_37:3) Triglycerides Triacylglyceride (16:0_38:2) Triglycerides Triacylglyceride (16:0_38:3) Triglycerides

[0203] Differential Biomarker Signature ofBD and MDD

[0204] The biomarker-based model showed a cross-validated AUROC (±SD) of 0.71±0.12 (P=7.3xl0‘7) in differentiating BD from MDD in the discovery cohort, and 0.73±0.06

[0205] (P=4.9xl0‘15) in the validation cohort (see Figures 19 and 20 and Table 4).

[0206] Table 4 - Performance metrics of the biomarker-based model in the discovery and validation cohorts

[0207] The final model consisted of 17 biomarkers, among which ceramide (Cer) dl8:0 / 24: l was the most important (see Figures 21 to 39 and Table 5).

[0208] Table 5 - Metabolite summary statistics

[0209] Alanine Ala Amino acids 0.165 0.781 0.080 0.740

[0210] Asparagine Asn Amino acids 0.147 0.656 0.008 0.919

[0211] Aspartic Acid Asp Amino acids -0.019 0.709 0.061 0.798

[0212] Cysteine Cys Amino acids 0.152 0.818 0.111 0.909

[0213] Glutamic Acid Glu Amino acids 0.432 1.211 0.229 1.269

[0214] Glycine Gly Amino acids 0.300 0.883 0.162 0.885

[0215] Isoleucine lie Amino acids 0.244 0.865 0.078 0.794

[0216] Leucine Leu Amino acids 0.159 0.757 0.010 0.762

[0217] Phenylalanine Phe Amino acids 0.204 0.873 0.054 0.787

[0218] Proline Pro Amino acids 0.289 0.865 0.203 1.085

[0219] Serine Ser Amino acids 0.420 1.175 0.147 0.999

[0220] Threonine Thr Amino acids 0.244 0.876 0.090 0.843

[0221] Tyrosine Tyr Amino acids 0.247 1.019 0.074 0.951

[0222] Valine Vai Amino acids 0.125 0.731 0.036 0.748

[0223] Deoxycholic acid DCA Bile acids 0.394 1.337 0.207 0.978

[0224] Glycocholic acid GCA Bile acids 0.535 1.162 0.383 1.191 Glycolithocholic acid sulfate GLCAS Bile acids 0.292 0.950 0.222 0.930

[0225] Taurocholic acid TCA Bile acids 1.647 8.181 0.487 2.314 beta-Alanine beta-Ala Biogenic amines 0.270 0.791 0.170 0.921 gamma-Amino- butyric acid GABA Biogenic amines 0.160 0.769 0.264 3.652 Carbohydrates and

[0226] Hexose Hl related 0.281 1.278 0.205 1.183

[0227] Hippuric acid HipAcid Carboxylic acids -0.031 1.080 0.062 1.066

[0228] Lactic acid Lac Carboxylic acids 0.500 1.026 0.139 0.738

[0229] Ceramide

[0230] (dl6: 1 / 18:0) Cer(dl6: 1 / 18:0) Ceramides 0.119 0.932 0.147 0.863

[0231] Ceramide

[0232] (dl6: 1 / 22:0) Cer(d l6: 1 / 22:0) Ceramides -0.050 0.826 0.129 0.832

[0233] Ceramide

[0234] (dl6: 1 / 24:0) Cer(d l6: 1 / 24:0) Ceramides 0.050 0.761 0.115 0.867

[0235] Ceramide

[0236] (dl8: 1 / 16:0) Cer(dl Ceramides 0.048 1.094 0.095 0.895

[0237] Ceramide

[0238] (dl8: 1 / 18:0) Cer(dl Ceramides 0.100 0.988 0.084 0.771

[0239] Ceramide

[0240] (dl8: 1 / 20:0(OH)) Cer(dl Ceramides 0.057 0.916 -0.020 0.831

[0241] Ceramide

[0242] (dl8: 1 / 20:0) Cer(dl Ceramides 0.069 0.993 0.092 0.760

[0243] Ceramide

[0244] (dl8: 1 / 22:0) Cer(dl Ceramides 0.116 1.027 0.107 0.790

[0245] Ceramide

[0246] (dl8: 1 / 23:0) Cer(dl Ceramides 0.037 0.864 0.104 0.787

[0247] Ceramide

[0248] (dl8: 1 / 24:0) Cer(dl Ceramides 0.218 0.930 0.132 0.792

[0249] Ceramide

[0250] (dl8: 1 / 24: 1) Cer(dl Ceramides 0.224 1.006 -0.031 0.864

[0251] Ceramide

[0252] (dl8: 1 / 26:0) Cer(dl Ceramides 0.224 0.931 0.160 0.803

[0253] Ceramide

[0254] (dl8: 1 / 26: 1) Cer(dl Ceramides 0.104 0.897 0.008 0.906

[0255] Ceramide

[0256] (dl8:2 / 16:0) Cer(dl Ceramides 0.113 1.008 0.113 0.886

[0257] Ceramide

[0258] (dl8:2 / 18:0) Cer(dl Ceramides 0.022 0.852 0.129 0.848

[0259] Ceramide

[0260] (d 18:2 / 20:0) Cer(dl Ceramides 0.024 0.917 0.060 0.911

[0261] Ceramide

[0262] (d 18:2 / 22:0) Cer(dl Ceramides 0.035 1.022 0.036 0.922

[0263] Ceramide

[0264] (d 18:2 / 24:0) Cer(dl Ceramides 0.038 1.073 0.045 0.928

[0265] Ceramide

[0266] (dl8:2 / 24: 1) Cer(dl Ceramides 0.134 0.909 -0.029 0.840

[0267] Dihexosyl ceramide

[0268] (dl8: 1 / 14:0) Hex2C Ceramides -0.043 0.870 0.096 0.962

[0269] Dihexosyl ceramide

[0270] (dl8: 1 / 16:0) Hex2C Ceramides 0.191 1.095 0.137 0.923

[0271] Dihexosyl ceramide

[0272] (dl8: 1 / 18:0) Hex2C Ceramides 0.070 0.750 0.088 0.780

[0273] Dihexosyl ceramide

[0274] (dl8: 1 / 20:0) Hex2C Ceramides 0.065 0.917 0.171 0.898

[0275] Dihexosyl ceramide

[0276] (dl8: 1 / 22:0) Hex2C Ceramides 0.056 0.868 0.170 0.862

[0277] Dihexosyl ceramide

[0278] (dl8: 1 / 24:0) Hex2C Ceramides 0.175 0.876 0.211 1.002

[0279] Dihexosyl ceramide (dl8: 1 / 24: 1) Hex2C Ceramides 0.231 0.892 0.097 0.864 Dihexosyl ceramide (dl8: 1 / 26:0) Hex2C Ceramides 0.069 0.829 0.230 1.121 Dihyd roceramide (d 18:0 / 22:0) Cer(d l Ceramides 0.169 0.659 0.102 0.712 Dihyd roceramide (d 18:0 / 24:0) Cer(d l Ceramides 0.195 0.802 0.056 0.824 Dihyd roceramide (dl8:0 / 24: 1) Cer(d l Ceramides 0.250 0.895 -0.026 0.794 Hexosylceramide (dl6: 1 / 22:0) HexCe Ceramides -0.038 0.811 0.121 0.804 Hexosylceramide (dl6: 1 / 24:0) HexCe Ceramides 0.005 0.928 0.073 0.877 Hexosylceramide (dl8: 1 / 14:0) HexCe Ceramides 0.019 0.890 0.034 0.805 Hexosylceramide (dl8: 1 / 16:0) HexCe Ceramides 0.054 0.923 0.121 0.774 Hexosylceramide (dl8: 1 / 18:0) HexCe Ceramides 0.072 0.980 -0.072 0.840 Hexosylceramide (dl8: 1 / 20:0) HexCe Ceramides -0.002 0.817 0.145 0.900 Hexosylceramide (dl8: 1 / 22:0) HexCe Ceramides -0.037 0.920 0.226 0.841 Hexosylceramide (dl8: 1 / 23:0) HexCe Ceramides 0.146 0.963 0.198 0.876 Hexosylceramide (dl8: 1 / 24:0) HexCe Ceramides 0.098 0.903 0.144 0.753 Hexosylceramide (dl8: 1 / 24: 1) HexCe Ceramides 0.111 0.844 0.116 0.812 Hexosylceramide (dl8:2 / 16:0) HexCe Ceramides -0.029 0.801 0.125 0.941 Hexosylceramide (dl8:2 / 18:0) HexCe Ceramides 0.041 0.738 0.019 0.849 Hexosylceramide (d 18:2 / 22:0) HexCe Ceramides 0.017 0.741 0.196 0.879 Trihexosylceramide (dl8: 1 / 16:0) Hex3C Ceramides -0.062 0.898 0.039 0.734 Trihexosylceramide (dl8: 1 / 20:0) Hex3C Ceramides 0.025 0.914 0.137 0.901 Trihexosylceramide (dl8: 1 / 22:0) Hex3C Ceramides 0.070 1.060 0.078 0.770 Trihexosylceramide (dl8: 1 / 24: 1) Hex3C Ceramides 0.189 1.049 -0.004 0.778

[0280] Trihexosylceramide (dl8: 1 / 26: 1) Hex3Cer(d Ceramides 0.195 0.814 -0.002 0.841 Cholesteryl ester 14:0 CE(14:0) Cholesteryl esters 0.195 0.930 0.209 1.032

[0281] Cholesteryl ester 14: 1 CE(14: 1) Cholesteryl esters 0.144 0.873 0.184 0.874

[0282] Cholesteryl ester 15:0 CE(15:0) Cholesteryl esters 0.188 1.482 0.088 0.869

[0283] Cholesteryl ester 16:0 CE(16:0) Cholesteryl esters 0.066 0.965 0.150 0.897

[0284] Cholesteryl ester 17:0 CE(17:0) Cholesteryl esters 0.469 1.227 0.236 0.830

[0285] Cholesteryl ester 17: 1 CE(17: 1) Cholesteryl esters 0.513 1.513 0.223 0.910

[0286] Cholesteryl ester 18:0 CE(18:0) Cholesteryl esters 0.079 0.956 -0.047 1.188

[0287] Cholesteryl ester 18: 1 CE(18: 1) Cholesteryl esters 0.144 0.960 0.228 0.922

[0288] Cholesteryl ester 18:2 CE(18:2) Cholesteryl esters -0.023 0.906 0.208 0.989

[0289] Cholesteryl ester 18:3 CE(18:3) Cholesteryl esters 0.149 0.938 0.333 2.362

[0290] Cholesteryl ester 20:0 CE(20:0) Cholesteryl esters 0.131 0.881 0.253 1.049

[0291] Cholesteryl ester 20: 1 CE(20: l) Cholesteryl esters 0.196 0.832 0.290 1.019

[0292] Cholesteryl ester 22:5 CE(22:5) Cholesteryl esters 0.068 0.876 0.281 1.165

[0293] Cholesteryl ester 22:6 CE(22:6) Cholesteryl esters 0.105 0.871 0.279 1.130 p-Cresol sulfate p-Cresol-S Cresols 0.086 0.862 0.182 0.821

[0294] Arachidonic acid AA Fatty acids 0.237 0.973 0.185 0.889 Docosahexaenoic acid DHA Fatty acids 0.118 0.682 0.194 0.873

[0295] Eicosapentaenoic acid EPA Fatty acids 0.247 0.958 0.267 1.020

[0296] Octadecadienoate FA(18:2) Fatty acids 0.033 0.952 0.082 0.840

[0297] Octadecenoic acid FA(18: 1) Fatty acids 0.272 1.118 0.166 1.085

[0298] Indoleacetic acid 3-IAA Indoles and derivatives 0.537 1.557 0.053 0.927

[0299] Indolepropionic acid 3-IPA Indoles and derivatives 0.037 1.315 0.416 1.419

[0300] Indoxyl sulfate Ind-SO4 Indoles and derivatives 0.141 0.900 0.173 0.842 Lysophosphatidyl- choline a C14:0 lysoPC a C14:0 Lysophosphatidylcholines 0.205 0.828 0.167 0.752 Lysophosphatidyl- choline a C16:0 lysoPC a C16:0 Lysophosphatidylcholines 0.197 0.893 0.163 0.905 Lysophosphatidyl- choline a C 16 : 1 lysoPC a C16: 1 Lysophosphatidylcholines 0.398 1.285 0.177 0.927 Lysophosphatidyl- choline a C17:0 lysoPC a C17:0 Lysophosphatidylcholines 0.147 0.763 0.120 0.854 Lysophosphatidyl- choline a CIS : 0 lysoPC a C18:0 Lysophosphatidylcholines 0.169 0.903 0.224 0.995 Lysophosphatidyl- choline a C18: 1 lysoPC a CIS: 1 Lysophosphatidylcholines 0.186 1.045 0.086 0.853 Lysophosphatidyl- choline a CIS : 2 lysoPC a C1S:2 Lysophosphatidylcholines 0.223 0.794 0.159 0.746 Lysophosphatidyl- choline a C20:3 lysoPC a C20:3 Lysophosphatidylcholines 0.313 0.963 0.209 0.864 Lysophosphatidyl- choline a C20:4 lysoPC a C20:4 Lysophosphatidylcholines 0.190 0.934 0.106 0.802 Lysophosphatidyl- choline a C24:0 lysoPC a C24:0 Lysophosphatidylcholines 0.243 1.092 0.211 0.943 Lysophosphatidyl- choline a C26:0 lysoPC a C26:0 Lysophosphatidylcholines 0.406 1.457 0.445 1.498 Lysophosphatidyl- choline a C26: 1 lysoPC a C26: l Lysophosphatidylcholines 0.166 0.958 0.156 0.900 Lysophosphatidyl- choline a C28:0 lysoPC a C28:0 Lysophosphatidylcholines 0.295 1.197 0.253 1.036 Lysophosphatidyl- choline a C28: 1 lysoPC a C2S: 1 Lysophosphatidylcholines 0.221 1.278 0.318 1.163

[0301] Hypoxanthine Hypoxanthine Nucleobases and related 0.480 1.319 0.400 1.334 Phosphatidyl-choline aa C24:0 PC aa C24:0 Phosphatidylcholines 0.306 1.326 0.229 1.102 Phosphatidyl-choline aa C28: 1 PC aa C2S: 1 Phosphatidylcholines -0.023 0.853 0.120 0.884 Phosphatidyl-choline aa C30:0 PC aa C30:0 Phosphatidylcholines 0.186 0.774 0.159 0.751 Phosphatidyl-choline aa C32:0 PC aa C32:0 Phosphatidylcholines 0.123 0.692 0.043 0.792 Phosphatidyl-choline aa C32: 1 PC aa C32: 1 Phosphatidylcholines 0.545 1.264 0.230 0.840 Phosphatidyl-choline aa C32:2 PC aa C32:2 Phosphatidylcholines 0.161 0.883 0.201 0.853

[0302] Phosphatidyl-choline aa C32:3 Phosphatidylcholines 0.036 0.979 0.203 0.881

[0303] Phosphatidyl-choline aa C34: l Phosphatidylcholines 0.147 0.987 0.050 0.771

[0304] Phosphatidyl-choline aa C34:2 Phosphatidylcholines 0.090 0.807 0.075 0.692

[0305] Phosphatidyl-choline aa C34:3 Phosphatidylcholines 0.175 0.935 0.129 0.713

[0306] Phosphatidyl-choline aa C34:4 Phosphatidylcholines 0.206 1.030 0.152 0.792

[0307] Phosphatidyl-choline aa C36:0 Phosphatidylcholines 0.090 0.803 0.097 0.689

[0308] Phosphatidyl-choline aa C36: 1 Phosphatidylcholines 0.120 0.943 0.101 0.802

[0309] Phosphatidyl-choline aa C36:2 Phosphatidylcholines 0.140 0.838 0.157 0.781

[0310] Phosphatidyl-choline aa C36:3 Phosphatidylcholines 0.171 0.770 0.171 0.716

[0311] Phosphatidyl-choline aa C36:4 Phosphatidylcholines 0.120 0.898 0.061 0.746

[0312] Phosphatidyl-choline aa C36: 5 Phosphatidylcholines 0.222 1.050 0.166 0.820

[0313] Phosphatidyl-choline aa C36:6 Phosphatidylcholines 0.195 0.955 0.134 0.771

[0314] Phosphatidyl-choline aa C38:0 Phosphatidylcholines 0.038 0.818 0.118 0.793

[0315] Phosphatidyl-choline aa C38:3 Phosphatidylcholines 0.069 0.839 0.125 0.887

[0316] Phosphatidyl-choline aa C38:4 Phosphatidylcholines 0.174 0.815 0.115 0.728

[0317] Phosphatidyl-choline aa C38: 5 Phosphatidylcholines 0.100 0.977 0.061 0.705

[0318] Phosphatidyl-choline aa C38:6 Phosphatidylcholines 0.062 0.957 0.094 0.862

[0319] Phosphatidyl-choline aa C40:2 Phosphatidylcholines 0.116 0.881 0.067 0.795

[0320] Phosphatidyl-choline aa C40:3 Phosphatidylcholines 0.088 0.827 0.065 0.844

[0321] Phosphatidyl-choline aa C40:4 Phosphatidylcholines 0.116 0.935 0.105 0.776

[0322] Phosphatidyl-choline aa C40: 5 Phosphatidylcholines 0.123 0.889 0.127 0.757

[0323] Phosphatidyl-choline aa C40:6 Phosphatidylcholines 0.058 0.917 0.084 0.809

[0324] Phosphatidyl-choline aa C42:0 Phosphatidylcholines 0.204 0.842 0.083 0.797

[0325] Phosphatidyl-choline aa C42: l Phosphatidylcholines 0.085 0.933 0.015 0.798

[0326] Phosphatidyl-choline aa C42:2 Phosphatidylcholines 0.013 0.912 0.045 0.771

[0327] Phosphatidyl-choline aa C42:4 Phosphatidylcholines 0.117 0.933 0.081 0.812

[0328] Phosphatidyl-choline aa C42: 5 Phosphatidylcholines 0.038 0.873 0.113 0.817

[0329] Phosphatidyl-choline aa C42:6 Phosphatidylcholines 0.180 0.947 0.106 0.792

[0330] Phosphatidyl-choline ae C30:0 Phosphatidylcholines 0.170 1.035 0.172 0.897

[0331] Phosphatidyl-choline ae C30: 1 Phosphatidylcholines 0.170 1.041 0.158 0.806

[0332] Phosphatidyl-choline ae C30:2 Phosphatidylcholines 0.010 0.862 0.109 0.774

[0333] Phosphatidyl-choline ae C32: 1 Phosphatidylcholines 0.012 0.863 0.011 0.731

[0334] Phosphatidyl-choline ae C32:2 Phosphatidylcholines -0.023 0.886 0.072 0.769

[0335] Phosphatidyl-choline ae C34:0 Phosphatidylcholines 0.174 0.840 0.076 0.774

[0336] Phosphatidyl-choline ae C34: l Phosphatidylcholines 0.001 0.894 -0.019 0.795

[0337] Phosphatidyl-choline ae C34:2 Phosphatidylcholines 0.026 0.797 0.067 0.699

[0338] Phosphatidyl-choline ae C34:3 Phosphatidylcholines 0.078 0.757 0.072 0.716

[0339] Phosphatidyl-choline ae C36:0 Phosphatidylcholines 0.191 0.834 0.082 0.713

[0340] Phosphatidyl-choline ae C36: 1 Phosphatidylcholines 0.080 0.901 0.061 0.665

[0341] Phosphatidyl-choline ae C36:2 Phosphatidylcholines 0.050 0.818 0.053 0.673

[0342] Phosphatidyl-choline ae C36:3 Phosphatidylcholines 0.055 0.872 0.061 0.725

[0343] Phosphatidyl-choline ae C36:4 Phosphatidylcholines 0.094 0.828 0.080 0.738

[0344] Phosphatidyl-choline ae C36: 5 Phosphatidylcholines 0.281 0.902 0.086 0.760

[0345] Phosphatidyl-choline ae C38:0 Phosphatidylcholines 0.089 0.705 0.092 0.683

[0346] Phosphatidyl-choline ae C3S : 1 Phosphatidylcholines 0.135 0.906 0.172 0.793

[0347] Phosphatidyl-choline ae C38:2 Phosphatidylcholines 0.088 0.874 0.123 0.758

[0348] Phosphatidyl-choline ae C38:3 Phosphatidylcholines 0.075 0.876 0.116 0.721

[0349] Phosphatidyl-choline ae C38:4 Phosphatidylcholines 0.039 0.828 0.077 0.714

[0350] Phosphatidyl-choline ae C38: 5 Phosphatidylcholines 0.134 0.885 0.067 0.735

[0351] Phosphatidyl-choline ae C38:6 Phosphatidylcholines 0.199 0.972 0.061 0.750

[0352] Phosphatidyl-choline ae C40: l Phosphatidylcholines 0.093 0.917 0.072 0.854

[0353] Phosphatidyl-choline ae C40:2 Phosphatidylcholines 0.017 0.980 0.044 0.801

[0354] Phosphatidyl-choline ae C40:3 Phosphatidylcholines 0.116 0.948 0.104 0.769

[0355] Phosphatidyl-choline ae C40:4 Phosphatidylcholines 0.123 0.996 0.114 0.742

[0356] Phosphatidyl-choline ae C40: 5 Phosphatidylcholines 0.114 1.010 0.085 0.758

[0357] Phosphatidyl-choline ae C40:6 Phosphatidylcholines 0.135 0.823 0.018 0.723

[0358] Phosphatidyl-choline ae C42: l Phosphatidylcholines 0.068 0.862 0.110 0.807

[0359] Phosphatidyl-choline ae C42:2 Phosphatidylcholines 0.079 0.847 0.028 0.768

[0360] Phosphatidyl-choline ae C42:3 Phosphatidylcholines -0.017 1.066 0.075 0.768

[0361] Phosphatidyl-choline ae C42:4 Phosphatidylcholines 0.091 0.907 0.078 0.735

[0362] Phosphatidyl-choline ae C44:3 Phosphatidylcholines 0.082 0.935 -0.010 0.825

[0363] Phosphatidyl-choline ae C44:4 Phosphatidylcholines 0.302 0.952 0.168 0.819

[0364] Phosphatidyl-choline ae C44: 5 Phosphatidylcholines 0.128 0.836 0.121 0.834

[0365] Phosphatidyl-choline ae C44:6 Phosphatidylcholines 0.131 0.892 0.099 0.751

[0366] Hydroxysphingomyelin C14: 1 Sphingomyelins -0.129 0.843 0.056 0.826

[0367] Hydroxysphingomyelin C16: 1 Sphingomyelins -0.163 0.744 0.038 0.757

[0368] Hydroxysphingomyelin C22: 1 Sphingomyelins -0.052 0.828 0.093 0.839 Hydroxysphingomyelin C22:2 Sphingomyelins -0.088 0.793 0.112 0.836 Hydroxysphingomyelin C24: 1 Sphingomyelins 0.077 0.753 0.033 0.667

[0369] Sphingomyelin C1 Sphingomyelins -0.096 0.955 0.003 0.840 Sphingomyelin C 1 Sphingomyelins -0.033 0.771 0.108 0.844 Sphingomyelin C1 Sphingomyelins -0.089 0.840 0.117 0.854 Sphingomyelin CI Sphingomyelins -0.055 0.703 0.158 0.856 Sphingomyelin C2 Sphingomyelins -0.101 0.747 0.167 0.855 Sphingomyelin C2 Sphingomyelins 0.100 0.809 0.098 0.815 Sphingomyelin C2 Sphingomyelins 0.069 0.922 0.034 0.857 Sphingomyelin C2 Sphingomyelins 0.057 0.901 0.058 0.909 Sphingomyelin C2 Sphingomyelins 0.144 0.745 0.014 0.831 Triacylglyceride (14:0_34: 1) Triglycerides 0.382 1.068 0.320 0.842 Triacylglyceride (14:0_36:2) Triglycerides 0.247 0.760 0.240 0.834 Triacylglyceride (14:0_36:3) Triglycerides 0.229 0.790 0.347 1.133 Triacylglyceride (16:0_30:2) Triglycerides 0.365 1.149 0.320 1.116 Triacylglyceride (16:0_32: 1) Triglycerides 0.569 1.295 0.379 0.863 Triacylglyceride (16:0_32:2) Triglycerides 0.525 1.366 0.369 0.920 Triacylglyceride (16:0_33: 1) Triglycerides 0.464 1.251 0.290 0.886 Triacylglyceride (16:0_34: 1) Triglycerides 0.439 1.009 0.294 0.865 Triacylglyceride (16:0_34:2) Triglycerides 0.464 1.185 0.300 0.916 Triacylglyceride (16:0_34:3) Triglycerides 0.385 1.013 0.283 0.924 Triacylglyceride (16:0_35: 1) Triglycerides 0.444 1.063 0.263 0.828 Triacylglyceride (16:0_35:2) Triglycerides 0.351 0.840 0.228 0.756

[0370] Triacylglyceride (16:0_35:3) TG(16:0_35:3) Triglycerides 0.215 1.157 0.196 0.854 Triacylglyceride (16:0_36:2) TG(16:0_36:2) Triglycerides 0.297 0.808 0.221 0.872 Triacylglyceride (16:0_36:3) TG(16:0_36:3) Triglycerides 0.184 0.823 0.167 0.929 Triacylglyceride (16:0_36:4) TG(16:0_36:4) Triglycerides 0.159 0.945 0.218 1.021 Triacylglyceride (16:0_37:3) TG(16:0_37:3) Triglycerides 0.256 0.756 0.193 0.898 Triacylglyceride (16:0_3S:2) TG(16:0_38:2) Triglycerides 0.261 0.776 0.205 0.926 Triacylglyceride (16:0_3S:3) TG(16:0_38:3) Triglycerides 0.232 0.746 0.207 0.960 Triacylglyceride (16:0_3S:4) TG(16:0_38:4) Triglycerides 0.051 0.965 0.087 0.940 Triacylglyceride (16: 1_32: 1) TG(16: 1_32: 1) Triglycerides 0.579 1.646 0.336 1.043 Triacylglyceride (16: l_34:0) TG(16: l_34:0) Triglycerides 0.427 1.387 0.182 0.945 Triacylglyceride (16: 1_34: 1) TG(16: 1_34: 1) Triglycerides 0.488 1.209 0.279 0.915 Triacylglyceride (16: 1_34:2) TG(16: 1_34:2) Triglycerides 0.398 1.199 0.283 1.110 Triacylglyceride (16: 1_36: 1) TG(16: 1_36: 1) Triglycerides 0.426 0.981 0.257 0.854 Triacylglyceride (16: 1_36:2) TG(16: 1_36:2) Triglycerides 0.342 1.092 0.169 0.954 Triacylglyceride (16: 1_36:3) TG(16: 1_36:3) Triglycerides 0.200 0.903 0.193 1.020 Triacylglyceride (16: 1_38:3) TG(16: 1_38:3) Triglycerides 0.101 0.997 0.120 1.229 Triacylglyceride (17:0_34: 1) TG(17:0_34: 1) Triglycerides 0.409 1.006 0.229 0.776 Triacylglyceride (17:0_34:2) TG(17:0_34:2) Triglycerides 0.266 1.241 0.170 0.889 Triacylglyceride (17: 1_34: 1) TG(17: 1_34: 1) Triglycerides 0.330 0.896 0.218 0.821 Triacylglyceride (17: 1_36:3) TG(17: 1_36:3) Triglycerides 0.014 0.908 0.137 0.930 Triacylglyceride (17:2_34:2) TG(17:2_34:2) Triglycerides 0.004 0.902 -0.010 0.878 Triacylglyceride (18:0_32: 1) TG(18:0_32: 1) Triglycerides 0.412 1.202 0.278 0.946

[0371] Triacylglyceride (1S:O_32:2) Triglycerides 0.240 1.041 0.170 0.882 Triacylglyceride (18:0_34:2) Triglycerides 0.360 1.117 0.204 0.906 Triacylglyceride (18:0_34:3) Triglycerides 0.229 0.965 0.148 0.912 Triacylglyceride (18: l_26:0) Triglycerides 0.336 1.435 0.361 1.414 Triacylglyceride (18: l_30:0) Triglycerides 0.458 1.164 0.359 0.863 Triacylglyceride (18: l_30: 1) Triglycerides 0.378 1.165 0.372 1.063 Triacylglyceride (18: l_30:2) Triglycerides 0.288 0.851 0.331 1.349 Triacylglyceride (18: l_32:0) Triglycerides 0.445 0.977 0.301 0.851 Triacylglyceride (18: 1_32: 1) Triglycerides 0.421 1.023 0.270 0.850 Triacylglyceride (18: 1_32:2) Triglycerides 0.265 0.890 0.272 1.002 Triacylglyceride (18: 1_32:3) Triglycerides 0.265 0.917 0.213 1.009 Triacylglyceride (18: l_33:0) Triglycerides 0.288 0.855 0.252 0.884 Triacylglyceride (18: 1_33: 1) Triglycerides 0.316 0.842 0.207 0.750 Triacylglyceride (18: 1_33:2) Triglycerides 0.174 0.831 0.187 0.808 Triacylglyceride (18: 1_34: 1) Triglycerides 0.306 0.856 0.211 0.878 Triacylglyceride (18: 1_34:2) Triglycerides 0.274 0.880 0.190 0.913 Triacylglyceride (18: 1_34:3) Triglycerides 0.182 0.884 0.208 0.978 Triacylglyceride (18: 1_35:2) Triglycerides 0.215 0.808 0.183 0.783 Triacylglyceride (18: 1_35:3) Triglycerides 0.227 0.766 0.209 0.905 Triacylglyceride (18: 1_36: 1) Triglycerides 0.352 0.941 0.347 1.378 Triacylglyceride (18: 1_38:5) Triglycerides -0.160 1.068 0.046 0.848 Triacylglyceride (18:2_30:0) Triglycerides 0.351 1.101 0.288 0.993

[0372]

[0373] Metabolite summary statistics. Shown are mean ± standard deviation values of robust-scaled analyte concentrations in the BD and MDD groups. Abbreviations: BD - bipolar disorder, MDD - major depressive disorder, SD - standard deviation.

[0374] Figures 22 to 38 show biomarker concentrations in the BD and MDD groups.

[0375] Specifically, robust-scaled analyte concentrations in the BD (n=67) and MDD (n = 174) groups from the original cohort are shown. The box plots show the interquartile range with the median marked by a bold horizontal line and whiskers extending to the minimum and maximum values (within 1.5 x interquartile range).

[0376] Figure 39 shows SHAP analysis of top biomarkers. The impact of biomarker predictors (vertical axis, ordered by importance) on the biomarker-based model output (horizontal axis) is shown. SHAP values above zero indicate a higher likelihood of a diagnosis of BD, and SHAP values below zero indicate a higher likelihood of a diagnosis of MDD. Individual data points represent SHAP values from the test sets, averaged across the 10 repeats of cross-validation, and coloured by standardized feature value.

[0377] Likewise, the most relevant analyte class was the ceramides. Figure 40 shows biomarker class importance. Cumulative contribution of compound classes to the overall accuracy of the biomarker-based model is shown. Cumulative importance was calculated by summing the importances of individual biomarkers within each compound category and averaging them across cross-validated models.

[0378] Added Predictive Value of Biomarkers

[0379] Combining biomarker readouts with patient-reported data led to significant improvements in the performance of diagnostic models based on demographic information, PHQ-9 scores, and MDQ outcomes (AAUROC±SD=0.05±0.08, 0.09±0.14, and 0.03±0.05, P=0.033, 0.026, and 0.028, respectively; see Figure 41 and Table 6).

[0380] Table 6 - Comparison of models based on patient self-reported endpoints with and without biomarker data The relative contribution of biomarkers to individual models varied from 8.8% to 99.9% (see Figure 42), with biomarker importance largely consistent across the models (see Figure 43).

[0381] Figure 42 shows the relative contribution of patient-reported symptom and demographic features and biomarker data to diagnostic model accuracy. Average cumulative importances from cross-validated models combing both questionnaire and biomarker data are shown. MDQ, Mood Disorder Questionnaire; PHQ-9, Patient Health Questionnaire-9; WEMBWS, Warwick-Edinburgh Mental Wellbeing Scale. Figure 43 shows biomarker importance in individual models. Average importance values from the cross-validated models combining patient self-reported information with biomarker data are shown. Clinical Utility of Biomarkers

[0382] Decision curve analysis showed that at relevant diagnostic thresholds, the inclusion of biomarkers may result in the additional identification of up to 30% of BD patients, accounting for false positives (P=1.5x l0‘4; see Figures 44-46 and Table 7)

[0383] Table 7 - Estimated clinical benefit of models based on patient self-reported endpoints with and without biomarker data

[0384] Figures 45 (original cohort) and 46 (validation cohort) show decision curve analysis of the biomarker-based model. The curves depict the standardized net benefit (vertical axis) across a range of diagnostic thresholds (horizontal axis) in the original (n=67 BD and 174 MDD patients) and validation cohorts (n=9 BD and 21 MDD patients). The horizontal and diagonal grey lines represent the benefit when assuming none or all patients have BD, respectively. The blue line corresponds to the null model, i.e. no information available. The vertical dashed line indicates the diagnostic cut-off determined using Youden's index. Values were averaged across the cross-validated models. Correlation with Psychopathology

[0385] The identified biomarkers correlated primarily with lifetime manic symptoms and psychiatric history (see Figure 47 and Table 8).

[0386] Table 8 - Correlation between biomarkers and psychopathology

[0387] Discussion

[0388] Differential diagnosis of BD and MDD in patients presenting with depressive symptoms poses a substantial clinical challenge, and there are currently no objective measures available to facilitate the diagnostic process. In the present study, the applicant identified and validated a metabolomic biomarker signature in patient DBS samples that can effectively distinguish BD from MDD during periods of low mood.

[0389] Furthermore, the potential clinical utility of the biomarker panel when combined with routinely collected patient information was demonstrated. The findings indicate that biomarkers may be a valuable source of clinically relevant information, and carry significant implications for the advancement of biomarker research of mood disorders.

[0390] It has been found that the incorporation of biomarkers enhances the discriminative ability of diagnostic models based on self-reported patient data, leading to a potential improvement in clinical utility. By themselves, the identified biomarkers demonstrated fair diagnostic performance (Li et al. (2018)) with an AUROC of 0.71 (0.73 in the validation cohort). This result outperformed all but one comparable study that employed high-content biological data to differentiate BD from MDD (Powell et al. (2014); Brunkhorst-Kanaan et al. (2019); Chen et al. (2020); Kittel-Schneider et a / . (2020); Idemoto et al. (2021); Liebers et al. (2021); Tomasik et a / . (2021); Li et al. (2022)), with the highest AUROC of 0.78 previously reported in a study utilizing gut microbiome data (Li et al. (2022)). When combined with patient-reported information, the added value of biomarkers was particularly evident in scenarios where data on manic symptoms were unavailable, such as when determining a diagnosis based on comprehensive demographic data or depressive symptoms.

[0391] Although smaller, significant improvements were also observed for the model based on the outcomes form the Mood Disorder Questionnaire (MDQ).

[0392] The finding that biomarkers primarily enhance diagnostic performance when information on manic symptoms is unavailable is consistent with the observation that the identified biomarkers correlated most strongly with these symptoms, and thus may, to some extent, constitute their surrogate markers. Importantly, the added value of biomarkers was observed at intermediate diagnostic thresholds, suggesting that they can be particularly helpful for patients whose diagnosis is uncertain. Furthermore, the relative importance analysis demonstrated that biomarkers substantially contributed to all models, accounting for up to 99.9% of the total feature importance in the model based on WEMWBS. Even in models where biomarkers did not provide clinical advantages, such as those based on psychiatric comorbidities or personality traits, they still contributed between 9% and 55% of the overall importance. This suggests that biomarkers may potentially replace patient selfreported information with a more objective measure while maintaining comparable levels of performance.

[0393] The current findings also suggest that the ceramide system may play a crucial role in the distinct mechanisms of mood disorders. Ceramides, which are precursors to all complex sphingolipids (Hannun & Obeid (2008); Mencarelli & Martinez-Martinez (2013)), have diverse physiological functions, including structural regulation of cell membranes and modulation of various cellular processes such as cell cycle and death, inflammation, and stress response (Stith et al. (2019)). The present work has identified elevated Cer(dl8:0 / 24: 1) as the most robust differential biomarker of BD and MDD. Previous lipidomic analyses of mood disorders have reported increased levels of circulating ceramides in both MDD (Gracia-Garcia et al. (2011); Brunkhorst- Kanaan et al. (2019); Tkachev et al. (2023)) and BD (Brunkhorst-Kanaan et al. (2019); Tkachev et al. (2023)) compared to healthy controls. However, these studies did not measure this specific analyte, nor directly compare BD to MDD (Tkachev et al. (2023)) or found differences between the disorders (Brunkhorst-Kanaan et al. (2019)).

[0394] Other research has suggested that the elevated ceramide levels in MDD may result from increased activity of acid sphingomyelinase (Kornhuber et al. (2005)), an enzyme that breaks down sphingomyelin into ceramide and phosphorylcholine. The expression of a transcript encoding an active form of this enzyme has also been found increased in MDD compared to healthy controls (Rhein et al. (2017)), causing a shift towards reduced sphingomyelin and increased ceramides levels. In this study, not only were increased levels of Cer(dl8:0 / 24: 1) in the BD group observed, but also decreased levels of sphingomyelins SM (OH) 014: 1 and SM C20:2, suggesting a potential role for acid sphingomyelinase in the distinct mechanisms of BD and MDD. Although some antidepressant drugs are known to target the acid sphingomyelinase-ceramide system (Gulbins et al. (2013)), the differences observed here are unlikely to be caused by antidepressant treatment, as the BD and MDD groups were matched for antidepressant use. This is further supported by the fact that the biomarker-based model showed a higher performance in discriminating BD from MDD compared to the model based on psychiatric history, which included extensive treatment data. These findings suggest that alterations in ceramide levels may not only differentiate BD and MDD from healthy controls, but also distinguish the conditions from each other.

[0395] In addition to ceramides and sphingomyelins, top biomarkers included also other lipids from the PC and TG classes as well as various non-lipid molecules. These biomarkers have not been compared directly between BD and MDD before, although some have been investigated independently for BD and MDD. For example, a recent large-scale lipidomic study by Tkachev et al. (2023) found higher circulating levels of PCs in MDD compared to healthy controls, which were absent in BD. However, the authors did not see consistent differences in TG levels, as observed in the present study.

[0396] In contrast, other studies have reported changes in TG levels in both BD (Naiberg et al. (2016); Huang et al. (2018); Dalkner et a / . (2021)) and MDD (van Reedt Dortland et a / . (2010); Han (2022)), which appeared to depend on the patients' affective states (van Reedt Dortland et a / . (2010); Huang et a / . (2018)) and may have been related to cardiometabolic comorbidities (Dalkner et al. (2021)) or treatment (Fjukstad et al. (2016)). However, both these factors were accounted for in the present analysis.

[0397] A practical aspect of the present work involves utilizing DBS samples and targeted metabolomics as means to identifying biomarkers of psychiatric conditions. DBSs provide several advantages over traditional blood sampling methods, such as being minimally invasive, enabling self-collection at home through a finger-prick, eliminating the need for centrifugation, the possibility of mailing through regular post, and allowing storage at room temperature. Although DBSs are a convenient and cost- effective method of blood sample collection, their routine clinical use is currently limited to neonatal disease screening (Nordfalk & Ekstrom (2019)).

[0398] In addition, metabolomic methods such as the one utilized in the current analysis offer several benefits over other commonly used biomarker discovery techniques, such as genomics and proteomics. The benefits include faster analysis times, low sample volume requirements, and standardized absolute quantification of hundreds of compounds representing all major biochemical classes, which ensures that the results are comparable across experiments, studies, and clinical sites. Furthermore, due to their highly dynamic nature, metabolites are considered to represent the complex interactions between genetic and environmental risk factors better than other omic data, and may therefore offer unique insights into disease states. The present applicant's findings suggest that DBS diagnostics offer a feasible tool to facilitate early diagnosis and treatment of mental health conditions (Barone et al. (2018); Han et al. (2020); Tomasik et al. (2021)).

[0399] Conclusions

[0400] This comprehensive diagnostic analysis identified a reproducible metabolomic biomarker signature in patients' DBSs which differentiated BD from MDD during episodes of low mood, and enhanced the predictive value of diagnostic models based on patients' self-reported information. The results highlight the potential of blood biomarkers to successfully complement psychometric assessments with more objective tools for diagnosing mood disorders, and point to the role of ceramides in the differential pathophysiology of BD and MDD.

[0401] Example 2 - Diagnosis of bipolar disorder

[0402] A dried blood spot sample is provided by a patient presenting with low mood. The levels of 17 biomarkers in the sample are determined and robust-scaled, and the values obtained are as follows:

[0403] Cer(dl8:0 / 24: 1) = 0.6837

[0404] TG(18: 1_38:5) = -0.3627

[0405] TrpBetaine = 0.9856

[0406] PC ae C36:5 = -0.4043

[0407] 3-IAA = 2.335

[0408] Lac = 0.9056

[0409] PC aa C32:3 = 0.4667

[0410] GCA = 2.274

[0411] Asn = 0.2928

[0412] TMAO = 0.6512

[0413] 3-IPA = -0.6995 t4-OH-Pro = 4.453

[0414] TG(17: 1_36:3) = 0.7643

[0415] SM (OH) C14: l = -0.08764

[0416] SM C20:2 = -0.3902

[0417] PC aa C42:0 = 1

[0418] HexCer(d 18: 1 / 22:0) = -0.637

[0419] These values are used to calculate the probability of having BD. The values are input into the machine learning model comprising of 19 extreme gradient-boosted trees (Figure 48).

[0420] The model converts the scaled biomarker concentration values into probabilities of having bipolar disorder, which are then summarised to give the final probability of the patient having BD (Figure 49):

[0421] The probability obtained is 0.715, and this is compared to the diagnostic threshold of 0.2956886. In this case the probability is higher than the threshold, and so a diagnosis of BD is made."

[0422] Example 3 - Diagnosis of major depressive disorder

[0423] A dried blood spot sample is provided by a patient presenting with low mood. The levels of 17 biomarkers in the sample are determined and robust-scaled, and the values obtained are as follows:

[0424] Cer(dl8:0 / 24: 1) = -0.9355

[0425] TG(18: 1_38:5) = 0.1815

[0426] TrpBetaine = 0.5025

[0427] PC ae C36:5 = -0.6301

[0428] 3-IAA = -0.2019

[0429] Lac = -0.6465

[0430] PC aa C32:3 = -0.08772

[0431] GCA = -0.3642

[0432] Asn = -0.9765 TMAO = 0.01094

[0433] 3-IPA = 0.5216 t4-OH-Pro = -0.5089

[0434] TG(17: 1_36:3) = -0.02564

[0435] SM (OH) C14: l = -0.2521

[0436] SM C20:2 = 0.7636

[0437] PC aa C42:0 = -1.053

[0438] HexCer(d 18: 1 / 22:0) = -0.6169

[0439] These values are used to calculate the probability of having BD. The values are input into the machine learning model comprising of 19 extreme gradient-boosted trees as for Example 2 (Figure 48).

[0440] The model converts the scaled biomarker concentration values into probabilities of having bipolar disorder, which are then summarised to give the final probability (Figure 50).

[0441] The probability obtained is 0.078, and this is compared to the diagnostic threshold of 0.2956886. In this case the probability is lower than the threshold, and so a diagnosis of MDD is made.

[0442] The disclosures in United Kingdom patent application 2313198.0, from which this application claims priority, and in the accompanying abstract are incorporated herein by reference.

[0443] List of References

[0444] • American Psychiatric Association (2013) Diagnostic and Statistical

[0445] • Manual of Mental Disorders, 5th Edn.

[0446] • Atluri et al. (2013) NeuroImage. Clin. 3, 123-31

[0447] • Barone et al. (2018) Front Psychiatry. 9, 636

[0448] • Benacek et a / . (2021) J Affect Disord . 295, 1122-1130

[0449] • Benacek et a / . (2022) Int J Bipolar Disord. 10, 15 • Bouckaert & Frank (2004) In: Dai H, Srikant R, Zhang C, eds. Advances in Knowledge Discovery and Data Mining. PAKDD 2004. Lecture Notes in Computer Science Springer; 3056, 3-12 Bourne et al. (2013) Acta Psychiatrica Scandinavica 128, 149-162. Bowden (2005) Journal of Affective Disorders 84, 117-125 Bragazzi (2013) Philos. Ethics. Humanit. Med. 8, 4 Brunkhorst-Kanaan et al. (2019) Metabolism 95, 65-76 Chen & Guestrin (2016) In: KDD '16: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 785-794 Chen et al. (2020) Front Psychiatry. 11, 757 Craddock & Sklar (2013) Lancet 381, 1654-62 Dalkner et al. (2021) Front Neurosci. 15, 717824 de Kluiver et al. (2023) Transl Psychiatry 13, 198 Einat (2007) Neuroscience & Biobehavioral Reviews 31, 850-857. Fried et al. (2014) Psychological Medicine 44, 2067-2076 Fjukstad et al. (2016) J. Clin. Psychopharmacol . 36, 615-620 GBD 2019 (2020) Lancet 396, 1204-1222

[0450] Ghaemi et al. (2000) Journal of Clinical Psychiatry 61, 804-808

[0451] Goldberg (1993) American Psychologist 48, 26-34

[0452] Gracia-Garcia et al. (2011) Journal of Neuropsychiatry and Clinical Neurosciences 23, 215-218

[0453] Gulbins et al. (2013) Nat. Med. 19, 934-938

[0454] Han (2022) Sci. Rep. 12, 6190

[0455] Han et al. (2020) Brain Behav. Immun. 90, 184-195

[0456] Hannun 8<. Obeid (2008) Nat. Rev. Mol. Cell. Biol. 9, 139-150

[0457] Haro et al. (2006) Int. J. Methods Psychiatr Res. 15, 167-180

[0458] Hirschfeld et al. (2000) Am J Psychiatry. 157, 1873-1875

[0459] Hirschfeld et al. (2003) The Journal of Clinical Psychiatry 64, 161-174

[0460] Hirschfeld et al. (2005) The Journal of the American Board of Family Medicine 18, 233-239

[0461] Holma et al. (2008) The Journal of Clinical Psychiatry 69, 1267-1275

[0462] Huang et al. (2018) Int. J. Psychiatry Med. 53, 273-281

[0463] Idemoto et al. (2021) J Psychiatr Res. 134, 48-56

[0464] Kessler & Ustun (2004) International Journal of Methods in Psychiatric Research 13, 93-121. Kittel-Schneider et al. (2020) Front Psychiatry. 11, 189

[0465] Kessler et al. (2005) Arch. Gen. Psychiatry 62, 617-627

[0466] Kornhuber et a / . (2005) J. Neural Transm. 112, 1583-1590

[0467] Kroenke et al. (2001) J Gen Intern Med. 16, 606-613

[0468] Kupfer et al. (2012) Lancet 379, 1045-1055

[0469] Li et al. (2018) Shanghai Arch Psychiatry. 30, 207-212

[0470] Li et al. (2022) J Affect Disord. 317, 166-175

[0471] Liebers et al. (2021) Psychol Med. 51, 1451-1458

[0472] Lundberg & Lee (2017) In: Proceedings of the 31st International Conference on Neural Information Processing Systems 4768-4777.

[0473] Martin-Key et al. (2021a) Brain and Behavior 11, e02167

[0474] Martin-Key (2021b) JMIR Form Res. 5, e27908

[0475] McCrone et al. (2008) British Journal of Psychiatry 20, 7-8

[0476] Mencarelli & Martinez-Martinez (2013) Cellular and Molecular Life Sciences. 70, 181-203

[0477] Merikangas et al. (2011). Archives of General Psychiatry 68, 241-251.

[0478] Mirea et al. (2021) JMIR Ment Health 8, e23813

[0479] Mitchell et al. (2009) Lancet 374, 609-619.

[0480] Nadeau & Bengio (2003) Mach. Learn. 52, 239-281

[0481] Naiberg et al. (2016) Acta Psychiatr Scand. 134, 241-248

[0482] Nordfalk & Ekstrom (2019) European Journal of Human Genetics 27, 203-210

[0483] Olbert et al. (2014) Journal of Abnormal Psychology 123, 452-462

[0484] Olmert et al. (2020) Journal of Medical Internet Research Research Protocols 9, el8453

[0485] Olesen et al. (2012) Eur. J. Neurol. 19, 155-62

[0486] Olmert et al. (2020) JMIR Research Protocols 9, el8453

[0487] Osby et al. (2016) The European Journal of Public Health 26, 867-871.

[0488] Otte et al. (2016) Nature Reviews Disease Primers 2, 16065

[0489] Passos (2016) Acta Psychiatrica Scandinavica 134, 91-103.

[0490] Paykel et al. (2008) Br. J. Psychiatry 193, 86-87

[0491] Powell et al. (2014) PLoS One 9; e91076

[0492] Regeer et al. (2015) International Journal of Bipolar Disorders 3, 22

[0493] Rhein et al. (2017) J. Affect. Disord. 209, 10-15

[0494] Schwarz & Bahn (2008) Electrophoresis 29, 2884-2890

[0495] Schwarz et al. (2009) In: Biomarkers for Psychiatric Disorders 75-96 (2009) • Scott et al. (2022) Acta Psychiatr Scand. 146, 389-405

[0496] • Stith et a / . (2019) 60, 913-918

[0497] • Tennant et al. (2007) Health Qual. Life Outcomes 5, 63

[0498] • Tkachev et al. (2023) JAMA Psychiatry 80, 250-259 • Tomasik et a / . (2021) Transl Psychiatry. 11, 41

[0499] • van Reedt Dortland et al. (2010) Journal of Clinical Psychiatry 71, 729-736

[0500] • Vickers & Elkin (2006) Medical Decision Making 26, 565-574

[0501] • Wittchen et al. (2011) Eur. Neuropsychopharmacol. 21, 655-79

[0502] • World Medical Association (2013) JAMA 310, 2191-2194 • WO 2007 / 003343

[0503] • WO 2007 / 003344

[0504] . WO 2016 / 1135750

[0505] • Youden (1950) Cancer 3, 32-35

Claims

CLAIMS1. Use of at least one of:Cer(dl8:0 / 24: 1)TG(18: 1_38:5)TrpBetainePC ae C36:53-IAALacPC aa C32:3GCAAsnTMAO3-IPA t4-OH-ProTG(17: 1_36:3)SM (OH) C14: lSM C20:2PC aa C42:0HexCer(d 18: 1 / 22:0) as a biomarker for the diagnosis of bipolar disorder or a predisposition to bipolar disorder, or for prognosing the development of bipolar disorder, or for monitoring efficacy of a therapy in an individual having, suspected of having, or being predisposed to bipolar disorder.

2. A method of diagnosing bipolar disorder or a predisposition thereto or prognosing the development of bipolar disorder, the method including:(a) measuring the amount of at least one of the following biomarkers in a biological sample from an individual:Cer(dl8:0 / 24: 1)TG(18: 1_38:5)TrpBetainePC ae C36:53-IAALacPC aa C32:3GCAAsnTMAO3-IPA t4-OH-ProTG(17: 1_36:3)SM (OH) C14: lSM C20:2PC aa C42:0HexCer(d 18: 1 / 22:0)(b) comparing the amount of the biomarker in the individual to a threshold value or to the amount of the biomarker in a control; wherein a difference in the amount of the biomarker is indicative of bipolar disorder or a predisposition thereto.

3. A method of monitoring efficacy of a therapy in an individual having, suspected of having, or being predisposed to bipolar disorder, the method including: measuring the amount of at least one of the following biomarkers in a biological sample from the individual:Cer(dl8:0 / 24: 1)TG(18: 1_38:5)TrpBetainePC ae C36:53-IAALacPC aa C32:3GCAAsnTMAO3-IPA t4-OH-ProTG(17: 1_36:3)SM (OH) C14: lSM C20:2PC aa C42:0HexCer(dl8: 1 / 22:0), wherein: a) the method is carried out on samples taken on two or more occasions from the individual and the amount of the biomarker present in the two or more samples is compared; and / or b) the amount of the biomarker present in the sample from the individual is compared with a threshold value or to one or more controls.

4. A method as claimed in claim 3, including comparing the amount of the biomarker in a sample obtained from the individual with the amount present in one or more samples taken from the individual prior to commencement of therapy, and / or one or more samples taken from the individual at an earlier stage of therapy.

5. A method as claimed in claim 3 or 4, wherein samples are taken prior to and / or during and / or following therapy for bipolar disorder.

6. Use as claimed in claim 1 in a method as claimed in any of claims 2 to 5.

7. A method or use as claimed in any preceding claim, including comparing the amount of the biomarker in a sample from the individual with the amount of the biomarker present in a sample from a subject not having bipolar disorder.

8. A method or use as claimed in any preceding claim, including measuring the amount of at least one of the following biomarkers in a biological sample from an individual:Cer(dl8:0 / 24: 1)TG(18: 1_38:5)TrpBetainePC ae C36:53-IAALacPC aa C32:3GCA.

9. A method or use as claimed in any preceding claim, including measuring the amount of at least Cer(dl8:0 / 24: 1) and / or 3-IAA.

10. A method or use as claimed in any preceding claim, including measuring at least two of the biomarkers.

11. A method or use as claimed in any preceding claim, including measuring the amount of Cer(dl8:0 / 24: 1) and at least one of:TG(18: 1_38:5)TrpBetainePC ae C36:53-IAALacPC aa C32:3GCAAsnTMAO3-IPA t4-OH-ProTG(17: 1_36:3)PC aa C42:0HexCer(dl8: 1 / 22:0).

12. A method or use as claimed in any preceding claim, including measuring the amount of Cer(dl8:0 / 24: 1) and at least one of:TG(18: 1_38:5)TrpBetainePC ae C36:5LacPC aa C32:3GCA.

13. A method or use as claimed in any preceding claim, including calculating an overall probability of the individual having bipolar disorder based on probabilities determined using each biomarker measured, and comparing the overall probability of the individual having bipolar disorder with a threshold probability or to a probability determined from a control.

14. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1) and TG(18: 1_38:5).

15. A method or use as claimed in any preceding claim, including measuring at least three of the biomarkers.

16. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5) and TrpBetaine.

17. A method or use as claimed in any preceding claim, including measuring at least four of the biomarkers.

18. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine and PC ae C36:5.

19. A method or use as claimed in any preceding claim, including measuring at least five of the biomarkers.

20. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5 and 3-IAA.

21. A method or use as claimed in any preceding claim, including measuring at least six of the biomarkers.

22. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA and Lac.

23. A method or use as claimed in any preceding claim, including measuring at least seven of the biomarkers.

24. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac and PC aa C32:

325. A method or use as claimed in any preceding claim, including measuring at least eight of the biomarkers.

26. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3 and GCA27. A method or use as claimed in any preceding claim, including measuring at least 10 of the biomarkers.

28. A method or use as claimed in any preceding claim, including measuring at least 15 of the biomarkers.

29. A method or use as claimed in any preceding claim, including measuring at least Cer(dl8:0 / 24: 1), TG(18: 1_38:5), TrpBetaine, PC ae C36:5, 3-IAA, Lac, PC aa C32:3, GCA Asn, TMAO, 3-IPA, t4-OH-Pro, TG(17: 1_36:3), SM (OH) C14: l, SM C20:2, PC aa C42:0 and HexCer(dl8: 1 / 22:0).

30. A method or use as claimed in any preceding claim, wherein the individual is suffering with or has suffered with low mood.

31. A method or use as claimed in any preceding claim, wherein the individual has previously been diagnosed with a mood disorder.

32. A method or use as claimed in any preceding claim, wherein the mood disorder is major depressive disorder.

33. A method or use as claimed in any preceding claim, wherein the biological sample is blood.

34. A method or use as claimed in any preceding claim, wherein the blood is in the form of a dried blood spot.

35. A method or use as claimed in any preceding claim, wherein quantifying is performed by measuring the concentration of the biomarker in the or each sample.

36. A method or use as claimed in any preceding claim, including the step of prescribing and / or administering a bipolar disorder medicament to the individual.

37. A method of treating a bipolar disorder patient identified by the use of a biomarker as claimed in any preceding claim or using the method of any preceding claim, including prescribing and / or administering a bipolar disorder medicament to said patient.

Images