Method for determining the likelihood of mental disorders using blood lipids

Abstract

The invention relates to the field of medicine, and more particularly to determining the likelihood of mental disorders in a person with the aid of blood lipids. The claimed method for determining the likelihood of a mental disorder in a person is based on a quantitative assay of a panel of lipids in a biological sample, particularly blood. This method has high sensitivity and specificity.

Description

[0001] Method for determining the probability of mental disorders using blood lipids

[0002] Field of technology

[0003] The invention relates to the field of medicine, namely to determining the probability of mental disorders using lipids in the blood.

[0004] State of the art

[0005] Mental health is currently one of the most serious problems facing all countries, as at least every fourth person experiences such problems at some point in their lives. The prevalence of mental health disorders in the world is very high. According to WHO (2019), about 20 million people suffer from schizophrenia; 45 million - from bipolar affective disorders; 280 million - from depression (https: / / www.who.int / news-room / fact-sheets / detail / mental-disorders). According to WHO forecasts, by 2030, depression will become the leading cause of disability (https: / / www.who.int / ru / news-room / fact-sheets / detail / depression). Improving the effectiveness of diagnostic methods for identifying mental disorders, in particular schizophrenia, assessing the risk of development and course of the disease is one of the pressing problems of modern psychiatry.

[0006] Mental disorders are currently diagnosed based on behavioral symptoms. There are no biochemical diagnostic tests for diagnosing mental disorders. Existing tests for assessing risk of disease are inadequate for screening purposes. For schizophrenia, for example, existing tests based on genetic information either provide poor estimates of the likelihood of a particular person developing the disease or cover only a small percentage of schizophrenia cases and cannot be used to test for the majority of cases.

[0007] As the most abundant compounds in the central nervous system (CNS), lipids play important, yet often neglected, roles in brain function. These roles include regulation of membrane fluidity and permeability, vesicle formation and transport, retrograde signaling, modulation of neurotransmitter release, and neural plasticity [1, 2]. Changes in lipid composition have also been linked to brain dysfunction, including common psychiatric disorders such as schizophrenia (SCZ), bipolar disorder (BPD), and major depressive disorder (MDD).

[0008] Thus, the development of an effective biochemical diagnostic test for the diagnosis of mental disorders is a relevant and important task. Disclosure of the invention

[0009] The problem solved by the present invention is to develop a method for determining the probability of the presence or absence of a mental state based on the quantitative determination of a panel of lipids in a biological sample, in particular in blood.

[0010] The stated problem is solved by implementing the proposed method for determining the probability of a mental disorder in a person using a panel of lipids in a biological sample, which includes the following steps:

[0011] (a) collection and preparation of a human biological sample;

[0012] (b) quantitative analysis of lipids of a human biological sample followed by calculation of the logistic regression index (P): where (bi,b2..,bk) = B are the calculated coefficients of logistic regression,

[0013] (ii, i2.. ik) = I — normalized lipid intensities for the studied sample, k — number of lipids;

[0014] (c) having received the value of P, the probability of the presence or absence of a mental disorder in a person is determined, while with a value of P>0.5, an increased probability of the presence of a mental disorder in a person is established, with a value of P<0.5, a decreased probability of the presence of a mental disorder.

[0015] In particular embodiments of the invention, at step a) lipid extraction is carried out.

[0016] In particular embodiments of the invention, the biological sample is blood, blood plasma, a dried drop of blood, or blood serum.

[0017] In particular embodiments of the invention, the quantitative analysis is a chromatograph mass spectrometric analysis, direct input mass spectrometry.

[0018] In special cases, the value of P is used to determine the high probability of a person having a mental disorder (with a value of P>0.8) and the low probability or absence of a mental disorder in a person (with a value of P<0.2), while the P values ​​themselves are not equal to the probability level, but demonstrate changes in the risk level - increased or decreased relative to the population.

[0019] In particular embodiments of the invention, the lipids are triglycerides, acylcarnitines, cholesterol esters, glycerophosphocholines, glycerophosphoethanolamines, glycerophosphoinositols, ceramides, sphingomyelins and / or fatty acids. As a result of implementing the method according to the invention, the following technical results are achieved:

[0020] - a new and effective method for diagnosing mental status has been developed, based on the quantitative determination of a panel of lipids in a biological sample, in particular in blood;

[0021] - the proposed method has high sensitivity (true positive proportion), namely 83%, and specificity (true negative proportion), namely 95%, ensured by using a unique specific panel of 20-80 lipids in the blood;

[0022] - the developed method can be used as a supplement / confirmation of existing methods of diagnosing mental disorders, as well as an independent study;

[0023] - the developed method expands the arsenal of available tools for primary large-scale screening of patients for the purpose of identifying mental disorders.

[0024] According to the present invention, blood lipids are proposed as biomarkers of mental disorders. A multivariate model is proposed for predicting the status of a mental disorder, including a panel of lipid biomarkers. Moreover, an extensive verification of the prognostic multivariate model was carried out for the blood lipid biomarkers according to the invention, and high prognostic efficiency was demonstrated, which emphasizes the high practical applicability of the method.

[0025] First, the proposed technology shows high diagnostic ability in separating healthy individuals from patients with mental disorders, in particular schizophrenia. Unlike existing (genetic) tests, which usually provide an estimate of relative risk and the absolute risk of the disease remains low, the proposed technology aims to provide the possibility of screening for diseases, when a positive test result would indicate a high probability of the disease. Second, the proposed technology is unique in that it uses a blood lipid panel consisting of a number of chemically different lipids. Such blood lipids are not currently used in clinical tests. In particular, only aggregate lipid measurements, such as total triglyceride or total cholesterol, are usually used to diagnose diseases such as cardiovascular disease.Although some lipid panels have been proposed and put into practice in some countries, such as a panel of several types of ceramides for the detection of cardiovascular diseases [3, 4], this example remains exceptional and also consists of a small number of lipid compounds.

[0026] Detailed disclosure of the invention Definitions (terms)

[0027] For a better understanding of the present invention, some terms used in the present description of the invention are provided below. The following definitions apply herein unless otherwise explicitly stated.

[0028] In the description of this invention, the terms "includes" and "including" are interpreted to mean "includes, among other things." These terms are not intended to be construed as "consists only of."

[0029] The term "and / or" means one, more, or all of the listed elements.

[0030] Also here, listing numeric ranges by endpoints includes all numbers within that range.

[0031] The term "optional" or "optional" or "optionally" as used herein means that the subsequently described event or circumstance may, but need not, occur and that the description includes instances in which the event or circumstance occurs and instances in which it does not occur.

[0032] A "mental disorder" (functional mental disorder) is a medical condition in which there is a disturbance of mood, thinking, perception, the ability to cope with everyday stress, to behave adequately to a situation, to compare the correctness of one's impressions, judgments and behavior with signals received from external reality. These disturbances occur independently of damage to brain tissue, i.e. they are manifestations of impaired mental processes, but not of an apparent structural defect of the brain. Examples of mental disorders associated with disturbances of the dopamine and serotonin neurotransmitter systems include, but are not limited to, schizophrenia, including treatment-resistant forms of schizophrenia, bipolar disorder, and psychotic depression.

[0033] The term "quantitative" refers to mass spectrometry, particularly chromatograph mass spectrometry or direct injection mass spectrometry, or other methods in combination with mass spectrometry, and denotes a measured value that is proportional to the concentration level of a compound.

[0034] The term "biological specimen" ("biosample") refers to blood, blood plasma, dried blood spot (DBS), or blood serum obtained or isolated from a patient.

[0035] The term "biomarker" as used herein means a substance capable of indicating a disease state. In the context of the present invention, aimed at diagnosing a mental disorder, in particular schizophrenia, "biomarker" means a substance indicating the risk of having a disease. "Biomarker" includes lipids, the amount of which is increased or decreased in subjects suffering from a mental disorder, in particular schizophrenia, compared to normal healthy subjects. The term "prediction" as used herein refers to the detection that a person has a significantly increased probability of having a disease.

[0036] As used herein, the term "lipid" refers to lipids from the following lipid classes: triglycerides, acylcarnitines, cholesterol esters, glycerophosphocholines (including their lyso- and ester forms), glycerophosphoethanolamines (including their lyso- and ester forms), glycerophosphoinositols, ceramides, sphingomyelins, fatty acids.

[0037] More specifically, the liquids can be separated from the group that is not classified: CAR 18:2, CAR 18:1 , СЕ 18:3, СЕ 18:2, СЕ 20:4, СЕ 20:3, СЕ 22:6, СЕ 22:5, DAY 41 :1 , DAY 41 :0, DAY 42:2, DAY 42:1 , DAY 44:2, DAY 44:1 , DAY 46:4, DAY 46:3, DAY 46:2, DAY 46:1 , DAY 46:0, DAY

[0038] 47:3, DAY 47:2, DAY 47:1 , DAY 48:5, DAY 48:4, DAY 48:3, DAY 48: 1 , DAY 48:0, DAY 49:4, DAY

[0039] 49:3, DAY 49:1 , DAY 50:6, DAY 50:5, DAY 50:4, DAY 50:2, DAY 50:1 , DAY 50:0, DAY 51 :5, DAY

[0040] 51 :4, DAY 51 :3, DAY 51:2, DAY 52:7, DAY 51 :0, DAY 52:6, DAY 52:5, DAY 52:4, DAY 52:3, DAY

[0041] 52:2, DAY 52:1 , DAY 53:6, DAY 53:4, DAY 53:3, DAY 54:6, DAY 54:5, DAY 54:4, DAY 54:3, DAY

[0042] 55:8, TAG 55:7, TAG 55:6, TAG 55:5, TAG 56:9, TAG 56:7, TAG 56:5, TAG 57:8, TAG 58:10, TAG 58:9, TAG 58:8, TAG 58:3, FA 10:3, FA 10:2, FA 10:1, FA 12:2, FA 13:1 , FA 21:1 , FA 23:1 , FA 24:4, FA 24:3, FA 24:2, FA 24:1 , FA 25:3, FA 25:2, FA 25:1 , FA 25:0, FA 26:5, FA 26:4, FA 26:2, FA 27:3, FA 28:4, LPC 14:0, LPC 15:0, LPC 16:1 , LPC 16:0, LPC 17:0, dCer 32:1 , LPC 18:3, LPC 18:2, LPC 18:1 , LPC 18:0, dCer 34:2, LPC 19:0, dCer 34:1 , LPC 20:5, LPC 20:4, LPC 20:2, LPC 20: 1 , LPC 20:0, dCer 36:2, dCer 36: 1 , LPC 22:6, LPC 22:5, LPC 22:4, DAG 34:3, dCer 38:1 , DAG 34:1 , DAG 36:5, DAG 36:4, DAG 36:3, dCer 40:2, DAG 36:2, dCer 40:1 , DAG 36:1 , dCer 41 :1 , PE_P 34:2, PE_P 34:1 , dCer 42:4, dCer 42:3, dSM 30:1 , dCer 42:2, dCer 42:1 , PE 34:2, dCer 43:2, PE_P 36:4, dCer 43:1 , PE_P 36:3, PE_P 36:2, dSM 32:2, dCer 44:2, dSM 32:0, PE_P 37:4, PE 36:4, PE 36:3, PE 36:2, PE 36:1 , dSM 33:2, PE_P 38:6, dSM 33:1 , PE_P 38:5, PE_P 38:4, dSM 34:3, dSM 34:2, dSM 34:1 , PE 38:6, PC 30:1 , PE 38:5,PC 30:0, PE 38:4, PE 38:3, PE_P 40:7, dSM 35:2, PE_P 40:6, PC_P 32:1 , dSM 35:1 , PE_P 40:5, PC_P 32:0, PE_P 40:4, PC_O 32:0, dSM 36:4, dSM 36:3, dSM 36:2, PC 32:2, dSM 36:1 , PE 40:6, PC 32:1 , PE 40:5, PE 40:5, PC 32:0, PC_P 34:3, PC_P 34:2, dSM 37:2, PC 33:2, PC_O 34:2, PC_P 34:1 , PC 33:1 , PC_O 34:1 , PC 33:0, PC_O 34:0, PC 34:4, dSM 38:3, PC 34:3, PC 34:3, PC_P 35:2, dSM 38:2, PC 34:2, PC 34:1 , PC 34:0, PC_P 36:5, PC 35:4, PC_O 36:4, PC_O 36:4, PC_O 36:4, PC_O 36:4, PC 35:3, PC_P 36:2, dSM 39:2, PC 35:2, PC_O 36:2, PC_P 36:1 , dSM 39:1 , PC 35:1 , PC_O 36:1 , PC 36:5, PC_P 37:4, PC_O 37:4, dSM 40:3, PC 36:3, dSM 40:2, PC 36:2, PC_P 37:1 , PC 36:1 , PC_P 38:6, PC_P 38:5, PC_P 38:5, PC 37:5, PC_O 38:5, PC_P 38:4, PC_O 38:4, dSM 41 :3, PC_O 38:3, dSM 41 :2, PC 37:2, PC_O 38:2, dSM 42:5, dSM 42:4, PC 38:4, PC 38:4, dSM 42:3, PC 38:3, dSM 42:2, PC 38:2, PC 38:1 , PC_O 40:7, PC_P 40:5, PC_O 40:5, PC_O 40:5, PC 39:4, PC_P 40:3, PC_O 40:4, PC_O 40:3, dSM 43:2, dSM 44:5, PC 40:5,PC 40:5, dSM 44:4, PC 40:4, dSM 44:3, PC 40:3, PC 40:3, dSM 44:2, PC 40:2, PC_O 42:6, PC_P 42:5, PC_O 42:5, PC_P 42:3, PC_O 42:4, PC 42:9, PC 42:6, PC 42:5, PC 42:4, PC_O 44:8, PC_P 44:6, PC_O 44:7, PC_P 44:5, PC_P 44:3, PC_O 44:4, PC_O 46:8, PC_O 46:7, PC_O 46:6 PC_O 46:5.,

[0043] The names for lipid compounds are given according to the following notation. The lipid class is indicated by an abbreviation (shown below), the two numbers correspond to the number of carbon atoms in the fatty acid / sphingoid base chains (before the colon) and the number of double bonds in them (after the colon). Thus, for example, PC 40:4 corresponds to a lipid of the phosphatidylcholine class with fatty acid residues in which a total of 40 carbon atoms and 4 double bonds are present.

[0044] DAG diacylglycerol;

[0045] TAG triacylglycerol;

[0046] FA free fatty acid;

[0047] CAR acylcarnitine;

[0048] CE cholesterol ester;

[0049] PC phosphatidylcholine;

[0050] PC_O plasmanyl-phosphatidylcholine;

[0051] PC_P plasmanyl-phosphatidylcholine;

[0052] LPC lysophosphatidylcholine;

[0053] 1_PC_O plasmanyl-lysophosphatidylcholine;

[0054] 1_PC_P plasmanyl-lysophosphatidylcholine;

[0055] PE phosphatidylethanolamine;

[0056] PE_P plasmanyl-phosphatidylethanolamine;

[0057] LPE lysophosphatidylethanolamine; dCer ceramide (with two hydroxy groups in the composition); dSM sphingomyelin (with two hydroxy groups in the composition).

[0058] The term "panel" as used herein refers to a composition, array, or collection comprising one or more biomarkers (lipids). The term may also refer to a profile or index of intensity patterns of one or more biomarkers described herein.

[0059] The term "intensity" in this document means the value of a lipid signature obtained as a result of measurements by means of, in particular, chromatograph mass spectrometry and processing of the obtained signals by software. The intensity of a lipid signature is a value proportional to the concentration of the corresponding lipid.

[0060] Unless otherwise defined, technical and scientific terms in this document have standard meanings generally accepted in the scientific and technical literature.

[0061] Brief description of the figures. Figure 1. Distribution of disease prediction probabilities for 102 samples, including both healthy (control group) people and people with mental disorders (schizophrenia). The figure clearly shows a bimodal distribution of probability estimates, reflecting the high diagnostic potential of the test.

[0062] According to the method of the invention, lipids are extracted from a blood sample (in particular, blood plasma) of a subject and the content of a panel of 20-80 types of lipids from the following lipid classes is quantitatively determined: triglycerides, acylcarnitines, cholesterol esters, glycerophosphocholines (including their lyso- and ester forms), glycerophosphoethanolamines (including their lyso- and ester forms), ceramides, sphingomyelins, fatty acids. The quantitatively determined lipid intensity is used in a logistic regression model with predetermined coefficients, which gives the probability of the disease for a specific person.

[0063] Plasma was obtained from peripheral venous blood in the morning from subjects who had undergone an overnight fast. Plasma samples were collected in 4 mL Vacutainer tubes containing the chelating agent ethylenediaminetetraacetic acid (EDTA). Lipids were extracted using a tert-butyl methyl ether (MTBE)-based extraction protocol. For mass spectrometric analysis, samples were diluted 1:5 and 1:2 with acetonitrile:isopropanol (7:3 (v:v)) for positive and negative ionization modes, respectively. The liquid chromatography / mass spectrometry system consists of a Waters Acquity high-performance liquid chromatography (HPLC) system (Waters, Manchester, UK) and a Q Exactive orbitrap mass spectrometer (Thermo Fisher Scientific, USA) equipped with a heated electrospray ionization (HESI) probe. Mass spectra were obtained in both positive and negative ionization modes.Mass spectra were processed using the open source XCMS software for peak detection and alignment. Relative lipid intensities were aligned with reference samples.

[0064] Mass spectrometry is widely known to those skilled in the art as a powerful tool for analyzing, identifying and quantifying various types of molecules (classes of compounds). The method is based on the ionization of a molecule, resulting in the formation of a "molecular ion", on the basis of which analysis and identification are carried out by measuring the mass-to-charge ratio and the intensity of the ion current. Depending on the nature of the molecule to be detected (e.g., lipid or protein origin), the complexity of the sample (mixture or pure substance) and the task of the study (e.g., qualitative or quantitative analysis), a suitable ionization and mass spectrometry method is selected. For the purposes of the present invention and the quantitative determination of lipids in a biological sample, mass spectrometry is used, in particular chromatograph mass spectrometric analysis or direct injection mass spectrometry, or other methods in combination with mass spectrometry.Chromatography coupled with mass spectrometry is a powerful tool for the analysis and quantification of lipids in complex biosamples.

[0065] The probability of mental disorders is estimated using the following formula: where (bi,b2..,bk) = B are the calculated coefficients of logistic regression, (h ,i2. ) = I are the normalized lipid intensities for the sample under study, k is the number of lipids.

[0066] A cutoff value of 0.5 was defined to distinguish psychiatric disorders from healthy controls, i.e., if P > 0.5, then the predicted label was set to an increased probability of having a psychiatric disorder.

[0067] Also, at the value of P, a high probability of the presence of a mental disorder in a person is determined (at the value of P>0.8) and a low probability or absence of a mental disorder in a person (at the value of P<0.2), while the P values ​​themselves are not equal to the probability level, but demonstrate changes in the risk level - increased or decreased relative to the population.

[0068] The present inventors have established a panel of plasma lipid changes that differentiates patients with mental disorders from controls.

[0069] Experimental part

[0070] A previously collected cohort of blood plasma from individuals with mental disorders (schizophrenia, depression, bipolar disorder, etc.) and healthy individuals (hereinafter referred to as the database) was used to build the diagnostic model, with a small portion of samples from this cohort being used as “reference samples”. A database of lipid intensities measured in individuals with mental disorders and controls was used to build the predictive model. The test blood samples, with which the model was validated, were measured together with 20 control reference samples whose relative intensities were available in the database.For each lipid, its intensity was transformed by subtracting the mean intensities of 20 reference samples re-measured with the test samples and adding the mean intensities from the 20 reference sample database, resulting in normalized lipid intensities.

[0071] Plasma was obtained from peripheral venous blood in the morning from subjects who had undergone an overnight fast. Plasma samples were collected in 4 ml Vacutainer tubes containing the chelating agent ethylenediaminetetraacetic acid (EDTA). Lipids were extracted using a tert-butyl methyl ether (MTBE)-based extraction protocol. For mass spectrometric analysis, samples were diluted 1:5 and 1:2 with acetonitrile:isopropanol (7:3 (v:v)) for positive and negative ionization mode, respectively. The liquid chromatography–mass spectrometry system consisted of a Waters Acquity HPLC system (Waters, Manchester, UK) and a Q Exactive orbitrap mass spectrometer (Thermo Fisher Scientific, USA) equipped with a heated electrospray ionization (HESI) probe. Mass spectra were obtained in both positive and negative ionization modes.Mass spectra were processed using the open source XCMS software for peak detection and alignment. A set of 365 lipids were used as a target list for quantification from the following lipid classes: CAR, CE, TAG, FA, LPC, LPE, LPC_O, LPC_P, dCer, dSM, DAG, PE, PE_P, PC, PC_P, PC_O. Of the 365 lipids targeted for quantification, 277 were reproduced (abbreviations according to the list of designations given above in the text):.

[0072] CAR 18:2, CAR 18:1, СЕ 18:3, СЕ 18:2, СЕ 20:4, СЕ 20:3, СЕ 22:6, СЕ 22:5, DAY 41:1, DAY 41:0, DAY 42:2, DAY 42:1, DAY 44:2, DAY 44:1, DAY 46:4, DAY 46:3, DAY 46:2, DAY 46:1, DAY 46:0, DAY 47:3, DAY 47:2, DAY 47:1, DAY 48:5, DAY 48:4, DAY 48:3, DAY 48:1, DAY 48:0, DAY 49:4, DAY 49:3, DAY 49:1 , DAY 50:6, DAY 50:5, DAY 50:4, DAY 50:2, DAY 50:1 , DAY 50:0, DAY 51 :5, DAY 51 :4, DAY 51 :3, DAY 51 :2, DAY 52:7, DAY 51 :0, DAY 52:6, DAY 52:5, DAY 52:4, DAY 52:3, DAY 52:2, DAY 52:1 , DAY 53:6, DAY 53:4, DAY 53:3, DAY 54:6, DAY 54:5, DAY 54:4, DAY 54:3, DAY 55:8, DAY 55:7, DAY 55:6, DAY 55:5, DAY 56:9, DAY 56:7, DAY 56:5, DAY 57:8, DAY 58:10, DAY 58:9, DAY 58:8, DAY 58:3, FA 10:3, FA 10:2, FA 10:1, FA 12:2, FA 13:1, FA 21:1, FA 23:1, FA 24:4, FA 24:3, FA 24:2, FA 24:1, FA 25:3, FA 25:2, FA 25:1, FA 25:0, FA 26:5, FA 26:4, FA 26:2, FA 27:3, FA 28:4, LPC 14:0, LPC 15-0, LPC 16:1, LPC 16:0, LPC 17:0, dCer 32:1, LPC 18:3, LPC 18:2, LPC 18:1, LPC 18:0, dCer 34:2, LPC 19:0,dCer 34:1 , LPC 20:5, LPC 20:4, LPC 20:2, LPC 20:1 , LPC 20:0, dCer 36:2, dCer 36:1 , LPC 22:6, LPC 22:5, LPC 22:4, DAG 34:3, dCer 38:1 , DAG 34:1 , DAG 36:5, DAG 36:4, DAG 36:3, dCer 40:2, DAG 36:2, dCer 40:1 , DAG 36:1 , dCer 41 :1 , PE_P 34:2, PE_P 34:1 , dCer 42:4, dCer 42:3, dSM 30:1 , dCer 42:2, dCer 42:1 , PE 34:2, dCer 43:2, PE_P 36:4, dCer 43:1 , PE_P 36:3, PE_P 36:2, dSM 32:2, dCer 44:2, dSM 32:0, PE_P 37:4, PE 36:4, PE 36:3, PE 36:2, PE 36:1, dSM 33:2, PE_P 38:6, dSM 33:1, PE_P 38:5, PE_P 38:4, dSM 34:3, dSM 34:2, dSM 34:1 , PE 38:6, PC 30:1 , PE 38:5, PC 30:0, PE 38:4, PE 38:3, PE_P 40:7, dSM 35:2, PE_P 40:6, PC_P 32:1 , dSM 35:1 , PE_P 40:5, PC_P 32:0, PE_P 40:4, PC_O 32:0, dSM 36:4, dSM 36:3, dSM 36:2, PC 32:2, dSM 36:1 , PE 40:6, PC 32:1 , PE 40:5, PE 40:5, PC 32:0, PC_P 34:3, PC_P 34:2, dSM 37:2, PC 33:2, PC_O 34:2, PC_P 34:1 , PC 33:1 , PC_O 34:1 , PC 33:0, PC_O 34:0, PC 34:4, dSM 38:3, PC 34:3, PC 34:3, PC_P 35:2, dSM 38:2, PC 34:2, PC 34:1 , PC 34:0, PC_P 36:5,PC 35:4, PC_O 36:4, PC_O 36:4, PC_O 36:4, PC_O 36:4, PC 35:3, PC_P 36:2, dSM 39:2, PC 35:2, PC_O 36:2, PC_P 36:1 , dSM 39:1 , PC 35:1 , PC_O 36:1 , PC 36:5, PC_P 37:4, PC_O 37:4, dSM 40:3, PC 36:3, dSM 40:2, PC 36:2, PC_P 37:1 , PC 36:1 , PC_P 38:6, PC_P 38:5, PC_P 38:5, PC 37:5, PC_O 38:5, PC_P 38:4, PC_O 38:4, dSM 41 :3, PC_O 38:3, dSM 41 :2, PC 37:2, PC_O 38:2, dSM 42:5, dSM 42:4, PC 38:4, PC 38:4, dSM 42:3, PC 38:3, dSM 42:2, PC 38:2, PC 38:1 , PC_O 40:7, PC_P 40:5, PC_O 40:5, PC_O 40:5, PC 39:4, PC_P 40:3, PC_O 40:4, PC_O 40:3, dSM 43:2, dSM 44:5, PC 40:5, PC 40:5, dSM 44:4, PC 40:4, dSM 44:3, PC 40:3, PC 40:3, dSM 44:2, PC 40:2, PC_O 42:6, PC_P 42:5, PC_O 42:5, PC_P 42:3, PC_O 42:4, PC 42:9, PC 42:6, PC 42:5, PC 42:4, PC_O 44:8, PC_P 44:6, PC_O 44:7, PC_P 44:5, PC_P 44:3, PC_O 44:4, PC_O 46:8, PC_O 46:7, PC_O 46:6, PC_O 46:5.,

[0073] Using the lipid intensity database, a logistic regression model with L1 regularization was trained to separate mental disorder from control. The LinearRegression function from the linear_model package of sklearn, a Python programming language, was used for training, and the regularization coefficient was C=0.1.

[0074] The regularization coefficient is the addition of a certain penalty for large values ​​of the coefficients in a linear model. This prohibits too "sharp" bends and prevents overfitting. ( https: / / neerc. ifmo.ru / wiki / index. р р?1Ше=Регуразиация or https: / / yandex.rU / q / datascience / 10672717313 / ).

[0075] . The number of lipids involved in the model is determined by the regularization coefficient. The regularization coefficient is set empirically depending on the target level of prediction accuracy tested on the validation sample.

[0076] For example, it was found that for the model involving 74 lipids, the accuracy on the validation set was lower than the set target value (e.g. 90%), while for the model involving 49 lipids, the accuracy on the validation set was higher than this target value. In the model involving 74 lipids, the regularization coefficient was automatically set at C=0.16, and in order to move to the model involving 49 lipids, the regularization coefficient is selected and in this example it is C=0.1.

[0077] A similar selection of coefficients is applicable for other models with other lipid quantities based on the validation sample.

[0078] Thus, in particular, in one of the specific examples, 49 lipids were studied. More specifically, the following 49 lipids were analyzed: CAR 18:2, CE 18:3, CE 22:5, TAG 46:4, TAG 50:0, TAG 55:6, TAG 56:5, FA 10:1, FA 12:2, FA 13:1, FA 25:0, FA 27:3, LPC 18:0, dCer 34:2, LPC 19:0, dCer 36:2, dCer 36:1, LPC 22:5, PE 38:6, PE_P 40:6, PE_P 40:4, dSM 36:3, dSM 36:2, PC 32:1, PC 32:0, PC_O 34:2, PC 33:1 , PC 34:4, dSM 38:3, PC 34:3, PC_P 35:2, PC 35:2, PC_P 36:1 , PC_O 37:4, PC_P 37:1 , PC_P 38:4, PC_O 38:3, dSM 41:2, PC 38:1 , PC_O 40:3, PC 40:5, dSM 44:4, PC 40:3, dSM 44:2, PC_O 42:5, PC_P 44:6, PC_O 44:7, PC_P 44:3, PC_O 46:5.

[0079] The corresponding calculated coefficients are B = (-0.567, - 0.104, 0.023, 0.046, -0.035, 0.445, 0.221, 0.107, -0.439, -0.288, 0.089, -0.292, -0.018, 0.062, - 0.208, 0.611, 0.046 0.028, -0.242, -0.318, 0.321, 0.057, 0.033, 0.021, 0.016, -0.137, 0.069, - 0.494, 0.158, -0.009, -0.069, -0.160, 0.098, -0.077, 0.195, -0.137, -0.075, -0.141 -0.015, -0.111, 0.185, 0.103, 0.186, 0.000, 0.046, -0.075, 0.154, 0.136, 0.134). The probability of mental disorders was estimated using the following formula: where (bi,b2..,bk) = B are the calculated coefficients of logistic regression, (ii,i2.) = I are the normalized lipid intensities for the sample under study, in this case k = 49.

[0080] A cutoff value of 0.5 was defined to discriminate the probability of having a mental disorder from healthy controls, i.e., if P>0.5, then the predicted label was set to an increased probability of having a mental disorder.

[0081] To test the predictive model, plasma lipid intensity was assessed in 102 samples, of which 55 were healthy controls and 47 individuals with psychiatric disorder, including patients with schizophrenia and first-episode psychosis. A lipid-based predictive model was tested for this dataset (Fig. 1), resulting in a sensitivity (true positive rate) of 0.83 and a specificity (true negative rate) of 0.95.

[0082] Below are some examples of the obtained probability estimates for test samples.

[0083] One blood sample from a person with schizophrenia had the following normalized intensities:

[0084] I = (-1.868, 0.535, 0.597, 0.615, 1.026, 0.635, -1.061, -0.915, -1.277, 1.753, -1.687, -0.630, -0.830, -0.892, 0.587, -0.681 , -0.551 , -0.288, -0.907, -3.392, -1.950, -1.098, -0.655, 0.588, 0.566, -0.635, 0.826, 0.532, -0.135, -0.233, 0.526, 1.168, 1.478, 1.166, 1.317, -0.801, -0.801, -0.900 0.691, 1.756, 0.516, 0.616, -0.161, -0.937, 1.754, 0.402, -0.758, 1.793, 2.880). The predicted probability of mental disorder was 0.85, and the predicted label was “psychiatric disorder”.

[0085] One blood sample from a healthy control had the following normalized intensities:

[0086] I = (-0.320, -0.979, 0.181, -0.114, -0.904, -0.895, -1.122, 1.257, 1.135, 1.824, 1.662, - 1.666, 1.270, 0.584, 1.570, -0.212, , -0.142, 0.735, 1.917, 2.573, 1.071 , 1.237, 0.809, 0.194, 0.614, 0.817, 0.707, 0.644, 1.271 , 1.215, 1.268, 1.841 , 1.170, 1.366, -0.174, 1.409, -0.963, 2.475, 1.413, 0.948, 0.641, 1.008, -1.114, 1.348, -0.213, -0.209, 1.177, 0.118, 0.449). The predicted probability of mental disorder was 0.06, and the predicted label was “healthy control”.

[0087] One blood sample from an individual with a first psychotic episode had the following normalized intensities:

[0088] I = (0.090, 0.447, 1.874, -0.520, 0.515, 1.551, 1.484, -0.688, 0.197, 0.816, 2.113, -0.182, -0.457, 0.606, -0.015, 1.684, 1.880, -0.604, 0.651, -1.133, -0.021, 1.049, 1.983, 0.724, 1.004, - 2.979, 1.162, 0.963, 2.676, -0.539, 0.046, 0.163, 0.680, 0.435, 1.314, 0.741, -0.994, -0.805, -0.140, -0.470, 0.433, 0.011, 0.346, 1.703, 0.876, 0.177, -0.788, -0.647, 1.153). The predicted probability of having a mental disorder was 0.96, and the predicted label was high probability of having a mental disorder.

[0089] Other models were also constructed with a different number of biomarkers by setting a different regularization coefficient (in this case, C=0.16 and C=0.04).For example, for a model that involved 74 lipids (CAR 18:2, CAR 18:1, CE 18:3, CE 22:5, TAG 46:4, TAG 47:1, TAG 50:0, TAG 54:4, TAG 55:6, TAG 56:5, FA 10:3, FA 10:1, FA 12:2, FA 13:1, FA 23:1, FA 25:1, FA 25:0, FA 27:3, FA 28:4, LPC 15:0, dCer 32:1, LPC 18:0, dCer 34:2, LPC 19:0, dCer 36:2, dCer 36: 1, LPC 22:5, DAG 34:3, PE 34:2, dCer43:2, PE_P 37:4, dSM 34:3, PE 38:6, PE_P 40:6, PC_P 32:1, PE_P 40:4, dSM 36:3, PC 32:1, PC 32:0, PC_P 34:3, PC_O 34:2, PC 33:1, PC 33:0, PC 34:4, dSM 38:3, PC 34:3, PC_P 35:2, PC_P 36:5, PC 35:2, PC_P 36:1, dSM 39:1, PC_O 37:4, PC_P 37:1, PC_P 38:4, PC_O 38:4, PC_O 38:3, dSM 41:2, PC 37:2, PC_O 38:2, dSM 42:5, PC 38:1, PC_O 40:5, PC 39:4, PC_O 40:3, PC 40:5, dSM 44:4, PC 40:3, PC_O 42:6, PC_O 42:5, PC_O 44:8, PC_P 44:6, PC_O 44:7, PC_P 44:3, PC_O 46:5), the coefficients are:.

[0090] B=(-0.585, -0.026, -0.175, 0.062, 0.052, 0.066, -0.087, 0.050, 0.533, 0.219, 0.010, 0.139, -0.476, -0.354, -0.013, 0.033, 0.105, -0.357, -0.027, 0.059, 0.016, -0.145, 0.082, -0.211 , 0.603, 0.103, 0.066, 0.023, -0.050, -0.035, -0.064, -0.035, -0.298, -0.394, -0.041, 0.382, 0.135, 0.159, 0.006, -0.035, -0.087, 0.039, -0.000, -0.623, 0.238, -0.065, -0.057, 0.141 , -0.154, 0.182, -0.007, -0.097, 0.247, -0.140, -0.105, -0.173, -0.186, -0.044, 0.121, 0.022, -0.083, -0.016, -0.026, -0.311 , 0.126, 0.208, 0.346, -0.060, 0.204, -0.004, -0.182, 0.194, 0.266, 0.166).

[0091] For the same sample of the first episode of psychosis mentioned earlier, the following values ​​of normalized intensities were obtained: i=(0.093, 1.203, 0.451, 1.875, -0.523, 0.731, 0.509, 1.175, 1.548, 1.463, -0.887, -0.703, 0.192, 0.798, -0.100, 2.031, 2.107, -0.185, - 3.707, 0.546, 1.164, -0.454, 0.600, -0.011, 1.678, 1.865, -0.600, 1.658, -0.283, 3.010, 0.649, 0.895, 0.650, -1.136, -1.943, -0.014, 1.047, 0.719, 1.001, -2.337, -2.958, 1.158, 0.563, 0.957, 2.683, -0.549, 0.037, -0.478, 0.154, 0.677, 1.470, 0.430, 1.307, 0.729, 0.421, -0.994, -0.814, - 1.356, -4.282, 3.697, -0.133, 0.264, 1.961, -0.480, 0.431, 0.011, 0.351, -0.456, 0.880, -1.722, 0.177, -0.790, -0.650, 1.146) and the predicted probability of mental disorder was 0.98.For the model involving 24 lipids (CAR 18:2, TAG 55:6, TAG 56:5, FA 10:1 , FA 12:2, FA 13:1 , FA 27:3, LPC 19:0, dCer 36:2, PE 38:6, PE_P 40:6, PE_P 40:4, dSM 36:2, PC_O 34:2, PC 34:4, PC_P 35:2, PC 35:2, dSM 39:1 , PC_P 37:1 , dSM 41 :2, PC 37:2, PC 40:5, PC_O 44:4, PC_O 46:5), the coefficients are: B=(-0.451 , 0.033, 0.205, 0.002, -0.338, -0.122, -0.136, -0.025, 0.542, -0.032, -0.234, 0.187, 0.119, -0.167, -0.228, -0.159, -0.031, -0.004, 0.140, -0.061, -0.025, 0.214, 0.033, 0.082).

[0092] For the same sample, the corresponding values ​​of the normalized intensities were 1=(0.093, 1.548, 1.463, -0.703, 0.192, 0.798, -0.185, -0.011, 1.678, 0.650, -1.136, -0.014, 1.966, -2.958, 0.957, 0.037, 0.154, 1.470, 1.307, -0.814, -1.356, 0.431, 0.908, 1.146) and the predicted probability of mental disorder was 0.91.

[0093] Using the lipid intensity database, a logistic regression model was trained to separate the corresponding diagnostic diseases from the control. The LinearRegression function from the linear_model package of sklearn, a Python programming language, was used for training, and the regularization coefficient was C = 0.034 (for 20 lipids), 0.185 (for 80 lipids), and 0.05 (for 32 lipids).

[0094] For the model involving 20 lipids (CAR 18:2, TAG 56:5, FA 12:2, FA 13:1, FA 27:3, dCer 36:2, PE 38:6, PE_P 40:6, PE_P 40:4, dSM 36:2, PC_O 34:2, PC 34:4, PC_P 35:2, PC 35:2, PC_P 37:1, dSM 41:2, PC 37:2, PC 40:5, PC_O 44:4, PC_O 46:5), the estimated logistic regression coefficients are:

[0095] B=(-0.414, 0.183, -0.327, -0.100, -0.109, 0.500, -0.003, -0.232, 0.158, 0.109, -0.144, - 0.178, -0.163, -0.005, 0.126, -0.050, -0.060, 0.202, 0.015, 0.066).

[0096] For a blood sample of a patient with a first episode of psychosis (schizophrenia) (patient 1), the following values ​​of normalized intensities were obtained: i=(-0.957, 1.143, - 1.130, -0.550, -1.939, -0.071, 0.631, -1.095, 1.213, -1.698, -1.863, 0.921, 0.617, -1.446, 0.546, - 1.820, 0.186, 1.733, -0.053, 0.637) and the predicted probability of mental disorder was 0.88, which corresponds to the predicted label high probability of presence = mental disorder.

[0097] For the model involving 80 lipids (CAR 18:2, CAR 18:1, CE 18:3, CE 22:5, TAG 46:4, TAG 47:1, TAG 49:1, TAG 50:0, TAG 54:4, TAG 55:6, TAG 56:5, FA 10:3, FA 10:1, FA 12:2, FA 13:1, FA 23:1, FA 25:1, FA 25:0, FA 26:5, FA 27:3, FA 28:4, LPC 15:0, dCer 32:1, LPC 18:0, dCer 34:2, LPC 19:0, LPC 20:4, dCer 36:2, dCer 36:1, LPC 22:5, DAG 34:3, PE 34:2, dCer 43:2, PE_P 37:4, dSM 34:3, PE 38:6, PC 30:1, PE_P 40:6, PC_P 32:1, PE_P 40:4, dSM 36:3, PC 32:1, PC 32:0, PC_P 34:3, PC_O 34:2, PC 33:0, PC 34:4, dSM 38:3, PC 34:3, PC_P 35:2, PC_P 36:5, PC 35:4, PC 35:2, PC_P 36:1, dSM 39:1 ,PC_O 37:4, PC_P 37:1 , PC_P 38:4, PC_O 38:4, PC_O 38:3, dSM 41 :2, PC 37:2, PC_O 38:2, dSM 42:5, PC 38:1 , PC_O 40:5, PC 39:4, PC_O 40:3, PC 40:5, dSM 44:4, PC 40:3, PC_O 42:6, PC_O 42:5, PC 42:5, PC 42:4, PC_O 44:8, PC_P 44:6, PC_O 44:7, PC_P 44:3, PC_O 46:5), the estimated logistic regression coefficients are: B=(-0.554, -0.074, -0.193, 0.087, 0.051, 0.082, 0.008, -0.102, 0.052, 0.538, 0.237, 0.028, 0.151 , -0.484, -0.373, -0.039, 0.052, 0.110, 0.015, -0.371 , -0.045, 0.133, 0.047, -0.202, 0.102, - 0.228, -0.002, 0.597, 0.116, 0.081 , 0.047, -0.088, -0.056, -0.095, -0.088, -0.303, 0.015, -0.414, - 0.066, 0.408, 0.156, 0.233, 0.000, -0.050, -0.077, -0.013, -0.688, 0.270, -0.083, -0.050, 0.180, 0.000, -0.143, 0.228, -0.018, -0.089, 0.263, -0.129, -0.148, -0.200, -0.200, -0.047, 0.165, 0.033, -0.078, -0.043, -0.036, -0.371 , 0.106, 0.233, 0.398, -0.078, 0.262, 0.001 , -0.037, -0.013, -0.201 , 0.207, 0.302, 0.173).

[0098] For the blood sample of a patient with a first episode of psychosis (patient 1), the corresponding normalized intensity values ​​were: I = (-0.957, -0.356, 2.425, 0.632, 0.370, 2.120, 1.594, 0.386, -1.886, -0.271, 1.143, -3.609, -1.009, -1.130, -0.550, -4.073, 0.322, -0.312, 1.450, -1.939, -4.377, 1.016, 1.052, 0.294, -0.700, 0.184, 0.226, -0.071, 0.963, 0.359, -1.384, 0.780, -0.580, -0.006, -1.467, 0.631, 1.705, -1.095, -2.000, 1.213, -3.919, 2.364, -0.466, -1.020, -1.863, 1.214, 0.921, -6.677, -0.226, 0.617, 0.217, 0.364, -1.446, 0.739, - 0.555, 0.625, 0.546, -1.200, 0.005, 0.352, -1.820, 0.186, -2.005, -5.373, -0.260, 0.606, 2.620, - 1.270, 1.733, -2.717, 1.605, -0.236, 0.462, -0.071, 1.167, -1.174, -1.338, -0.946, -0.634, 0.637) and the predicted probability of mental disorder was 0.85, which corresponds to the predicted label high probability of having a mental disorder.

[0099] For the model involving 32 lipids (CAR 18:2, CE 18:2, TAG 55:6, TAG 56:5, FA 10:1, FA 12:2, FA 13:1, FA 25:0, FA 27:3, dCer 34:2, LPC 19:0, dCer 34:1, dCer 36:2, PE 38:6, PE_P 40:6, PE_P 40:4, dSM 36:2, PC 32:0, PC_O 34:2, PC 34:4, dSM 38:3, PC_P 35:2, PC 35:2, dSM 39:1, PC_P 37:1, PC_P 38:4, dSM 41:2, PC 40:5, PC 40:3, PC_O 44:7, PC_O 44:4, PC_O 46:5), the estimated logistic regression coefficients are:

[0100] B=(-0.489, -0.000, 0.142, 0.210, 0.033, -0.358, -0.172, 0.022, -0.176, 0.013, -0.060, 0.004, 0.585, -0.086, -0.250, 0.238, 0.098, 0.022, -0.175, -0.296, 0.043, -0.141, -0.075, -0.006, 0.151, - 0.022, -0.079, 0.221, 0.005, 0.027, 0.028, 0.115).

[0101] For the blood sample of a patient with a first episode of psychosis (patient 1), the corresponding values ​​of the normalized intensities were: 1 = (-0.957, -2.579, - 0.271, 1.143, -1.009, -1.130, -0.550, -0.312, -1.939, -0.700, 0.184, 0.301, -0.071, 0.631, -1.095, 1.213, -1.698, -0.466, -1.863, 0.921, -6.677, 0.617, -1.446, -0.555, 0.546, -1.200, -1.820, 1.733, 1.605, -0.946, -0.053, 0.637) and the predicted probability of mental disorder was 0.9, which corresponds to the predicted label of high probability of having a mental disorder.

[0102] For blood samples from a patient with depression (patient 2), the following normalized intensities were obtained:

[0103] 1=(-2.995, -0.859, -0.704, -1.644, -0.213, -0.073, 0.772, 0.008, 0.098, -0.228, 0.022, 0.843, 1.017, -0.079, 0.333, -0.083, -0.362, -0.081, 0.069, -0.003) (for the 20-lipid model above), l=(-2.995, -2.012, 1.307, -0.882, 1.396, 1.672, 1.196, 0.873, -0.845, -0.720, -0.859, -1.095, -1.800, -0.704, -1.644, 2.148, 0.833, 0.191, -2.238, -0.213, 0.318, -0.434, 1.189, 0.556, 0.253, - 0.496, -1.369, -0.073, 1.102, 0.398, -1.162, 2.426, -0.626, -0.566, -0.493, 0.772, 1.837, 0.008, 0.545, 0.098, -1.303, 1.344, 0.008, 1.280, 0.022, 0.243, 0.843, -2.043, -0.156, 1.017, 0.219, - 0.499, -0.079, 1.561, 0.675, -0.873, 0.333, -0.184, -0.968, 1.024, -0.083, -0.362, 1.551, 0.358, 0.315, -0.349, -1.135, 1.109, -0.081, -1.150, 0.763, -2.627, 0.068, -2.160, -0.597, -0.499, -1.806, 0.677, 0.464, -0.003) (for the 80 lipid model given above),

[0104] 1=(2.995, -0.315, -0.720, -0.859, -1.800, -0.704, -1.644, 0.191, -0.213, 0.253, -0.496, 1.016, -0.073, 0.772, 0.008, 0.098, -0.228, 0.008, 0.022, 0.843, -2.043, 1.017, -0.079, 0.675, 0.333, -0.184, -0.083, -0.081, 0.763, 0.677, 0.069, -0.003) (for the model of 32 lipids given in higher).

[0105] The resulting predicted probability of mental disorder for each of the 20, 80, and 32 lipid models was 0.77, 0.84, and 0.77, respectively, corresponding to the predicted label of increased probability of having a mental disorder, for the 20 and 32 lipid models and for the 80 lipid model corresponding to the predicted label of high probability of having a mental disorder.

[0106] For blood samples from a patient with bipolar disorder (patient 3), the following normalized intensities were obtained:

[0107] 1=(-1.295, -1.127, -0.005, -0.218, 0.085, 0.661, -0.017, 0.597, 0.781, 1.142, 0.454, -0.111, 0.905, 1.163, 1.334, 1.563, 0.333, 0.444, 0.562, 1.331) (for the 20-lipid model above),

[0108] 1=(-1.295, -0.386, 0.350, 0.831, -1.402, -0.471, -0.430, -0.459, -0.633, -0.757, -1.127, - 0.595, -0.135, -0.005, -0.218, -0.855, -0.049, 0.530, -1.210, 0.085, -0.068, -0.310, 0.434, -0.324, 0.433, -0.043, -0.993, 0.661, 1.273, -0.104, -1.144, -0.686, 0.833, 0.550, 0.673, -0.017, 0.325, 0.597, 0.458, 0.781, 0.897, 0.113, 0.564, 1.079, 0.454, 0.720, -0.111, 0.793, 0.022, 0.905, 0.675,

[0109] 0.139, 1.163, 1.294, 1.136, 1.003, 1.334, 0.399, 1.109, 1.788, 1.563, 0.333, 1.508, 0.086, 0.253,

[0110] 1.621, 0.988, 0.618, 0.444, 0.282, 0.250, 0.325, 1.759, 0.353, 0.175, 1.296, 0.136, 2.117, 1.526,

[0111] 1.331) (for the 80 lipid model given above),

[0112] 1=(-1.295, 1.135, -0.757, -1.127, -0.135, -0.005, -0.218, 0.530, 0.085, 0.433, -0.043, 1.047, 0.661, -0.017, 0.597, 0.781, 1.142, 0.564, 0.454, -0.111, 0.793, 0.905, 1.163, 1.136, 1.334, 0.399, 1.563, 0.444, 0.250, 2.117, 0.562, 1.331) (for a model of 32 lipids, given above).

[0113] The resulting predicted probability of mental disorder for each of the 20, 80 and 32 lipid models was: 0.7, 0.89 and 0.74, respectively, which corresponds to the predicted label of increased probability of having a mental disorder, for the 20 and 32 lipid models and for the 80 lipid model, the predicted label corresponds to high probability of having a mental disorder. Although the invention has been described with reference to the disclosed embodiments, it should be obvious to those skilled in the art that the specific experiments described in detail are provided only for the purpose of illustrating the present invention and should not be considered as limiting the scope of the invention in any way. It should be understood that various modifications can be made without departing from the essence of the present invention.

[0114] List of references, which are included in this description of the invention as references:

[0115] 1. Piomelli D, Astarita G, Rapaka R. A neuroscientist’s guide to lipidomics. Nat Rev Neurosci. 2007.

[0116] 2. Lauwers E, Goodchild R, Verstreken P. Membrane Lipids in Presynaptic Function and Disease. Neuron. 2016.

[0117] 3. Mantovani A, Dugo C. Ceramides and risk of major adverse cardiovascular events: A meta-analysis of longitudinal studies. J Clin Lipidol. 2020.

[0118] 4. Hilvo M, Vasile VC, Donato LJ, Hurme R, Laaksonen R. Ceramides and Ceramide Scores: Clinical Applications for Cardiometabolic Risk Stratification. Front Endocrinol (Lausanne). 2020.

Claims

Invention formula 1. A method for determining the probability of a mental disorder in a person using a panel of lipids in a biological sample, comprising the following steps: (a) collection and preparation of a human biological sample; (b) quantitative analysis of lipids of a human biological sample followed by calculation of the logistic regression index (P): where (bi,b2..,bk) = B are the calculated coefficients of logistic regression, (i1.i2-.ik) = I — normalized lipid intensities for the studied sample, k — amount of lipids; (c) having received the value of P, the probability of the presence or absence of a mental disorder in a person is determined, whereby a value of P>0.5 establishes an increased probability of the presence of a mental disorder in a person, while a value of P<0.5 establishes a decreased probability of the presence of a mental disorder.

2. The method according to claim 1, wherein in step a) lipids are extracted.

3. The method according to claim 1, wherein the biological sample is blood, blood plasma, dried blood spot or blood serum.

4. The method according to claim 1, wherein the quantitative analysis is a chromatograph mass spectrometric analysis, direct injection mass spectrometry.

5. The method according to claim 1, wherein the lipids are triglycerides, acylcarnitines, cholesterol esters, glycerophosphocholines, glycerophosphoethanolamines, glycerophosphoinositols, ceramides, sphingomyelins and / or fatty acids.

Images