Diagnosis of mood disorders using blood RNA editing biomarkers

Abstract

The present invention is directed to a method for in vitro diagnosis of a mood disorder, particularly a depressive disorder, more preferably a major depressive episode (DEP), in a human patient from a biological sample of the patient, using at least one or a combination of specific A to I-edited RNA biomarkers associated with specific depression scores and algorithms. The present invention also is directed to a method for monitoring treatment of a mood disorder, preferably for monitoring depression, in a human subject, which method implements the method for diagnosing a mood disorder of the present invention. A kit for determining whether a patient exhibits a mood disorder, preferably a depressive disorder, more preferably a DEP, is also included in the present invention.

Description

[Technical Field]

[0001] The present invention relates to a method for in vitro diagnosing mood disorders, particularly depressive disorders, and more preferably major depressive episodes (DEPs), in human patients from biological samples of said patients, using at least one or a combination of specific A to I edited RNA biomarkers associated with specific depression scores and algorithms. The present invention also relates to a method for monitoring the treatment of mood disorders, preferably depression, in human subjects, the method of which implements the method for diagnosing mood disorders of the present invention. Kits for determining whether a patient presents with a mood disorder, preferably depressive disorder, are also included in the present invention.

[0002] Major depressive episode (DEP) is a very common, heterogeneous stress-related mental disorder, the most prevalent, affecting approximately 10% of men and 20% of women worldwide. DEP is associated with a significant increase in mortality, primarily due to suicidal behavior, as well as an increased risk of developing other medical conditions, such as heart disease, diabetes, and stroke. A chronic and relapsing course of DEP, associated with treatment resistance, is frequently observed—the remission rate after the initial course of treatment is 37%, further increasing the use of the healthcare system. Currently, DEP is included in the Diagnostic and Statistics Manual for Mental Disorders. 1 Diagnosis is made through a clinical evaluation against the criteria outlined in the 5th edition of the Criteria for Depression and Psychiatry. DEP is characterized by an episode lasting at least two weeks, during which the individual experiences five or more distinct symptoms, such as a combination of depressed mood, loss of interest in daily life, sleep dysregulation, fatigue, or indecisiveness. The main problem, therefore, is that the diagnosis of DEP is based on descriptive criteria and is therefore inherently subjective. Accurate diagnosis is further complicated by the heterogeneous nature of DEP, which may not perfectly match the most common scales of depression, as well as by the similarities of symptoms with other mental and physical disorders.

[0003] Numerous studies have identified potential biomarkers for DEP that could help diagnose the disease or provide reliable alternatives to clinical responses to treatment. 2 This is being investigated. The definition of a biomarker is "a feature that is objectively measured and evaluated as an indicator of a normal biological process, pathogenic process, or pharmacological response to a therapeutic intervention." Thus, biomarkers can be used as diagnostic tools, for staging diseases, and / or for predicting and monitoring clinical responses to interventions. Developing biomarkers that may be suitable for diagnostic testing in the field of mood disorders is a complex and challenging task. Several biomarkers for DEP have been proposed, including circulating pro-inflammatory cytokines (e.g., IL-6, TNFα, IL-1β) or inflammatory markers CRP, neurotrophic factors (e.g., BDNF), and hormones, e.g., ACTH or cortisol (see discussion for further details). 3 (See [reference]). However, when used individually, none of these biomarkers meet the criteria for sensitivity and specificity for clinical use. Over the past decade, numerous studies have focused on brain biomarkers, including functional neuroimaging studies, or have investigated all ~omics fields, including genomics, transcriptomics, proteomics, or metabolomics.

[0004] Most promising methods appear to reside in multi-parameter biomarker panels for which a biological explanation exists. These biomarkers need to be quantified using standard blood samples and research methods, and in many cases, specific diagnostic algorithms are being developed in parallel. 4 A major challenge in identifying suitable biomarkers for DEP diagnostic assays is that it may require the use of a combination of several biomarkers that reflect changes in different biological mechanisms.

[0005] Accumulating evidence suggests that RNA editing, primarily carried out by the ADAR (adenosine deaminase acting on RNA) enzyme, may be absolutely involved in neurological or immunological disorders among other pathologies. Inflammation is known as a virulence factor that may explain part of the pathophysiology of depression. Recent evidence supports the concept that RNA editing plays a crucial role in the immune-inflammatory process. This shows that ADAR1 modulates important immune responses through RNA editing. Recent studies have reported an overall increase in RNA editing in SLE, a chronic autoimmune disorder that can greatly increase autoantigen loading and accelerate the progression of systemic lupus erythematosus (SLE). By editing specific primary miRNAs that target the IL-6 receptor subunit gp130, ADAR2 regulates gp130-dependent IL-6 signaling. Previously, alterations in RNA editing were demonstrated in psychotic patients with chronic hepatitis C virus (HCV) receiving IFN-α and ribavirin antiviral combination therapy, where treatment-induced depression is a common complication. RNA editing in the coding sequences of key mood-regulating proteins, including AMPA and kainate (KA) subtypes of glutamate and serotonin receptors in the CNS, can modulate the functional properties of encoded protein products. Reduced editing of the R / G site of the AMPA receptor, resulting from ADAR2 downregulation, may play a role in the pathophysiology of mental disorders. RNA editing at the Q / R site of AMPA, a glutamate ion channel transmembrane receptor, may be involved in the pathogenesis of mood disorders and the action of antidepressants. Alterations in serotonin receptors have been identified in the brains of patients with major depressive disorder or those who have attempted suicide due to depression. Recent studies have identified modifications in PDE8A RNA editing in the brains of offspring of suicide victims as a brain marker related to the immune response to suicide. 5 .

[0006] This study focused on whole-transcriptome RNA editing modifications by RNA-Seq in DEP patients to identify editing variants that can be used as biomarkers for clinically useful diagnostic assays of DEP. Using whole-transcriptome RNA editing modification measurements, significant differential editing ratios were identified in numerous genes associated with immune system and CNS function. Subsequently, the top 15 best biomarkers were selected for further validation in a prioritized cohort. These RNA editing variants were involved in different pathways associated with mood disorders and DEP, e.g., immune system or neurotransmitter receptors, and postsynaptic transmission. Machine learning approaches were applied to combine RNA editing variants and optimize predictions, and then the biomarker panel was validated in a large cohort of 411 depressed patients and age, race, and sex-matched controls. The RNA editing variant panel, with its high specificity and sensitivity, distinguished DEP patients from healthy controls. This study is the first to evaluate multiple RNA editing variants in blood samples from DEP patients. The research findings will contribute to a better understanding of the molecular pathophysiology of DEP and pave the way for the development of diagnostic assays for DEP with clinical applications.

[0007] In a first aspect, the present invention relates to a method for a selected at least one biomarker, preferably a combination of at least two biomarkers, which can be used to diagnose mood disorders, particularly depressive disorders, more preferably major depressive episodes, in human patients from a blood sample of said patient, wherein the method is a) A step of analyzing the RNA-Seq dataset using the Editome analysis pipeline to identify A-to-I differentially edited locations with a minimum coverage of 30x, ensuring high reliability at specific base locations, b) A step of conducting a differential analysis to identify areas where editing may differ specifically between patients with mood disorders, particularly depressive disorders, more preferably DEP, and healthy controls. c) applying a pre-specified quality criterion of the group consisting of coverage > 30, AUC > 0.6, 0.95 > fold change > 1.05, p < 0.05, and exclusion of intergenic regions; d) optionally, preferably, by calculating a global Alu editing index (AEI), which is a measure of the overall ratio of RNA editing of Alu repeats, to confirm that no difference in global RNA editing was observed between patients and controls; e) identifying and selecting biomarkers that reflect changes in different biological processes including immune and CNS (central nervous system) functions by GSEA (gene set enrichment analysis) of the biomarkers selected in step c) or d), preferably by using gene ontology tools; f) applying a specified quality criterion of the group consisting of coverage > 30, AUC > 0.8, 0.8 > fold change > 1.20, and p < 0.05) to a combination of several biomarkers selected in step e) that represent different biological mechanisms in two separate groups and clearly distinguish between healthy controls and patients with mood disorders, particularly depressive disorders, more preferably major depressive episodes.

[0008] In a preferred embodiment, in step f), the combination of at least two biomarkers is selected based on whether the combination is statistically significant with a p-value < 10 -4 between controls and cases.

[0009] In a second aspect, the present invention is an in vitro method for diagnosing mood disorders, particularly depressive disorders, more preferably DEP in a human patient from a biological sample of the patient, the method comprising a) for one A to I editing RNA biomarker, the relative proportion of RNA editing at a given editing site for at least one or a combination of sites within the RNA transcript of the t biomarker, and / or - determining the relative percentages of an isoform (including the edited site) or an unedited isoform, or a combination of isoforms of the RNA transcript of said biomarker, wherein said at least one A to I editing RNA biomarker is selected from the group consisting of PRKCB, MDM2, PIAS1, IFNAR1, IFNAR2, LYN, AHR, RASSF1, GAB2, CAMK1D, KCNJ15, IL17RA, and PTPRC biomarkers; b) determining a value resulting from the percentage obtained in step a); c) determining whether the resulting value or the depression score obtained in step b) is higher or not higher than a control value obtained for a healthy control subject, wherein the control value is determined in the same manner as the resulting value obtained in step b); classifying whether the depression score obtained for the patient is higher or not higher than a threshold (cut-off), and if the depression score for the patient is higher than the threshold, classifying the patient as positive or having a mood disorder, particularly a depressive disorder, more preferably DEP, or if the depression score is not higher than the threshold, classifying the patient as negative or not having a mood disorder, particularly a depressive disorder, more preferably DEP.

[0010] The reference symptoms of mood disorder, depressive disorder, and DEP are well known to those skilled in the art and are described, for example, in the DSM-5 (trademark) textbook from the American Psychiatric Association (2013, pages 155-188).

[0011] In the present invention, depressive disorders include disruptive mood dysregulation disorder, major depressive disorder (including major depressive episodes), persistent depressive disorder (dysthymia), premenstrual dysphoric disorder, substance / drug-induced depressive disorder, depressive disorder due to another medical condition, other specific depressive disorders, and unspecified depressive disorder.

[0012] In a preferred embodiment, the present invention provides an in vitro method for diagnosing mood disorders, preferably depressive disorders, more preferably DEP, in human patients from biological samples of said patients, wherein the method is a) For a combination of at least two A to I edited RNA biomarkers, - For each selected biomarker, the relative proportion of RNA editing at a given editing site to at least one or a combination of sites that can be edited in the RNA transcript of the at least two biomarkers, and / or - Determining the relative percentage of one isoform or combination of isoforms of each RNA transcript of the two biomarkers, The determination that at least two A to I edited RNA biomarkers are selected from the group consisting of PRKCB, MDM2, PIAS1, IFNAR1, IFNAR2, LYN, AHR, RASSF1, GAB2, CAMK1D, KCNJ15, IL17RA, PTPRC, and PDE8A biomarkers, b) Determining a final result value based on an algorithm or equation that includes the result values ​​obtained for each of the at least two biomarkers and the result values ​​obtained for each selected biomarker, c) Determining the final outcome values ​​or depression scores obtained for patients, and the control outcome values ​​obtained for healthy control subjects, wherein the control outcome values ​​are determined in a manner equivalent to that of the final outcome values. d) A method comprising: classifying the patient as positive or having a mood disorder, preferably a depressive disorder, more preferably DEP, if the final result value or depression score value is higher than a threshold / cutoff; or classifying the patient as negative or not having a mood disorder, preferably a depressive disorder, more preferably MDD, if the final result value is not higher than the threshold.

[0013] For example, but not limited to, in the diagnostic method according to the present invention, the threshold, cutoff, or depression score may be: -When using the mROC method, the Z-value calculated in healthy control patients, or - The probability that a patient is considered to have a mood disorder, depressive disorder, or DEP.

[0014] In this document, the terms biomarker and target have the same meaning and may be used interchangeably.

[0015] In a preferred embodiment, in a method for diagnosing depression according to the present invention, a combination of at least two biomarkers is such that the combination has a p-value between a healthy control subject and a case (a patient with a mood disorder, depressive disorder, or DEP) < 10-4 The selection is based on whether or not it is statistically significant.

[0016] In a more preferred embodiment, in step a), the A to I edited RNA biomarker is selected from the group consisting of PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, and GAB2.

[0017] In a more preferred embodiment, at least in step a), a combination of at least two A to I edited RNA biomarkers is used. The at least two A to I edited RNA biomarkers are selected from the group consisting of PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A biomarkers.

[0018] In a more preferred embodiment, the present invention provides an in vitro method for diagnosing mood disorders, preferably depressive disorders, and more preferably major depressive episodes. In step b), the result value or depression score is calculated by an algorithm that implements a multivariate method, and the multivariate method is - In particular, an mROC algorithm for maximizing AUC (Area Under Curve) ROC, providing an equation for each combination and identifying a linear combination that can be used as a new virtual marker Z, Z = a.(biomarker 1) + b.(biomarker 2) + ...i.(biomarker i) + ...n.(biomarker n) In the formula, a, b, i, n are calculated coefficients, and (biomarker 1, 2, i, n) is the level of the biomarker being considered (i.e., the level of the RNA editing site or isoform of a given target / biomarker), mROC algorithm and / or - A random forest (RF) algorithm, and / or arbitrarily applied to evaluate combinations of RNA editing sites and / or isoforms, particularly to rank the importance of RNA editing sites and / or isoforms, and to find the best combination of RNA editing sites and / or isoforms. - A multivariate analysis applied to evaluate combinations of RNA editing sites and / or isoforms for diagnostic purposes, wherein the multivariate analysis is, for example, - Logistic regression models that can be applied to univariate and multivariate analyses to estimate the relative risk of patients at different levels of RNA editing sites or isoform values. - The CART (Classification And Regression Tree) approach applied to evaluate combinations of RNA editing sites and / or isoforms. - Support Vector Machine (SVM) approach, - Artificial Neural Network (ANN) approach, - Bayesian network approach, - WKNN (Weighted k-Nearest Neighbors) approach, -Partial Least Squares Discriminant Analysis (PLS-DA), - Linear and quadratic discriminant analysis (LDA / QDA), as well as - The method includes multivariate analysis, selected from a group of other mathematical methods that combine biomarkers.

[0019] In a particularly preferred embodiment, the mood disorder or depressive disorder to be diagnosed is a major depressive episode (DEP).

[0020] In another preferred embodiment, the present invention relates to a method according to the present invention for diagnosing mood disorders, particularly depressive disorders, which include mild, moderate, and severe depressive disorders.

[0021] In another preferred embodiment, the present invention relates to a method according to the present invention for diagnosing mood disorders, particularly depressive disorders, wherein the depression can be classified as mild, moderate, or severe depression depending on the level of difference between the depression score value obtained in step b) and the control result value obtained for a control subject.

[0022] In a preferred embodiment, the method according to the present invention for diagnosing mood disorders, particularly depressive disorders, the biological sample is whole blood, serum, plasma, urine, or cerebrospinal fluid, with whole blood being the most preferred biological sample.

[0023] A selected biomarker is preferred in which the result value or depression score obtained in step c) is statistically / specifically different from the control result value at p<0.05.

[0024] Furthermore, regarding the biomarker or combination of biomarkers selected in step a), - Coverage > 30, -AUC>0.6, preferably AUC>0.8, -0.95 > Magnification change > 1.05, and The method of the present invention is preferred, provided that the criterion -p<0.05 is met.

[0025] Furthermore, the selected A to I edited RNA biomarker included in step a) a) The method of the present invention is preferred to be selected from the group consisting of combinations of at least 3, 4, 5, 6, 7, or 8 of the following biomarkers: PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A, preferably selected from the group consisting of combinations of the eight cited biomarkers.

[0026] Furthermore, when the model is adjusted for age, sex, psychiatric treatment, and addiction, or when the model is adjusted for age and sex, preferably when a random forest (RF) algorithm is used to evaluate combinations of RNA editing sites and / or isoforms, in particular to rank the importance of RNA editing sites and / or isoforms, and to combine the best RNA editing sites and / or isoforms, the method of the present invention is preferred, where the selected A to I edited RNA biomarkers include, in step a), a combination of the following eight biomarkers: GAB2, IFNAR1, KCNJ15, LYN, MDM2, PRKCB, CAMK1D, and PDE8A.

[0027] Furthermore, to evaluate combinations of RNA editing sites and / or isoforms, particularly to rank the importance of RNA editing sites and / or isoforms, and to combine the best RNA editing sites and / or isoforms, a random forest (RF) algorithm is applied, and when the model is adjusted for age, sex, psychiatric treatment, and addiction, or when the model is adjusted for age and sex, the selected A to I edited RNA biomarkers include combinations of the following eight biomarkers in step a): GAB2, IFNAR1, KCNJ15, LYN, MDM2, PRKCB, CAMK1D, and PDE8A, and preferably the calculation of the relative percentage of at least one of the RNA editing sites or isoforms is based on the RNA editing sites or isoforms listed in Tables 5.1 and 5.2, which is preferred in the present invention.

[0028] Furthermore, the method of the present invention is preferable in which the selected A to I edited RNA biomarker includes at least one or a combination of 2, 3, 4, 5, 6, 7, or 8 selected biomarkers in step a), and the calculation of the relative percentage of at least one RNA editing site or isoform is based on the RNA editing sites or isoforms listed in Table 2A (editing sites), 2B (isoforms), or 3 (editing sites) for each of the selected biomarkers.

[0029] Furthermore, the method of the present invention is preferable in which the combination of at least 2, 3, 4, 5, 6, 7, or 8 selected biomarkers is selected from the biomarker combinations shown in Tables 4 and 5, preferably combinations including the following eight biomarkers: PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A.

[0030] Furthermore, a combination is preferred that includes eight biomarkers: PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A, selected from the combinations listed in Table 6.

[0031] Furthermore, the method of the present invention is preferable in which the algorithm or equation enabling the calculation of the result value or depression score Z is selected from the Z equations listed in the examples, preferably equations that implement combinations of the following eight biomarkers: PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A.

[0032] In another particular embodiment, the present invention relates to an in vitro method according to the present invention for diagnosing mood disorders, particularly depressive disorders, more preferably MDD, in a human patient, wherein the patient being tested is female, and in step a) at least one or a combination of the at least two A to I edited RNA biomarkers is selected from the group of biomarkers consisting of PRKCB, MDM2, PIAS1, IFNAR1, IFNAR2, LYN, AHR, RASSF1, GAB2, CAMK1D, KCNJ15, IL17RA, and PTPRC, and if one biomarker is selected, or if a combination of at least two biomarkers (or targets) is selected, the group further comprises A to I edited RNA PDE8A.

[0033] Furthermore, the method of the present invention is preferred in which the combination of at least 2, 3, 4, 5, 6, 7, or 8 selected biomarkers is selected from the group consisting of PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A biomarkers (or targets), particularly the combination of biomarkers shown in the examples or figures in patients of the female population or the Z equation, in particular the combination including the eight biomarkers PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A, more preferably the combination listed in Table 7.

[0034] In another particular embodiment, the present invention relates to an in vitro method according to the present invention for diagnosing depression in a human patient, wherein the patient being tested is male, and in step a) at least one or a combination of the at least two A to I edited RNA biomarkers is selected from the group of biomarkers consisting of PRKCB, MDM2, PIAS1, IFNAR1, IFNAR2, LYN, AHR, RASSF1, GAB2, CAMK1D, KCNJ15, IL17RA, and PTPRC, and if one biomarker is selected, or if a combination of at least two biomarkers (or targets) is selected, the group further comprises A to I edited RNA PDE8A.

[0035] Furthermore, the method of the present invention is preferred in which the combination of at least 2, 3, 4, 5, 6, 7, or 8 selected biomarkers is selected from the group consisting of the PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A biomarkers (or targets), particularly the combination of biomarkers shown in the examples or figures in patients in the male population, or the Z equation, in particular the combination including the eight biomarkers PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A, more preferably the combination listed in Table 9.

[0036] In a more preferred embodiment, the present invention relates to an in vitro method according to the present invention for diagnosing depression in human patients, wherein the combination of biomarkers used in the diagnosis of mood disorders, particularly depressive disorders, and / or the Z-equation or algorithm associated with such combination differs depending on whether the patient being tested is female or male.

[0037] Preferably, the combination of biomarkers and / or Z-equation, depending on whether the patient being tested is female or male, is selected from the combinations and / or Z-equations shown in the examples or figures for this particular case.

[0038] Furthermore, in step a), the method of the present invention is preferable in which the relative proportion of RNA editing and / or the percentage of isoforms in a given edit are measured in the biological sample by NGS.

[0039] Furthermore, in step a), the method of the present invention is preferable in which the amplicon / nucleic acid sequence used to detect the RNA editing site and / or isoform of the RNA transcript of the biomarker is obtained or can be obtained with a set of primers (SEQ ID NOs: 1-36) listed in Table 1 for each of the selected biomarkers.

[0040] Furthermore, in step a), the method of the present invention is preferable in which the amplicon / nucleic acid sequence used to detect the RNA editing site and / or the isoform of the RNA transcript of the biomarker includes an amplicon or nucleic acid sequence obtainable by a set of primers (SEQ ID NOs: 1 to 36) listed in Table 1, or can be obtained by a set or combination of primers that enables obtaining the same amplicon or nucleic acid sequence.

[0041] The amplicon or nucleic acid sequence used to detect RNA editing sites and / or isoforms of the RNA transcript of the biomarker may be the complete RNA transcript sequence itself.

[0042] In another aspect, the present invention relates to a method for monitoring mood disorders, preferably depressive disorders, in human subjects from blood samples of said subjects, preferably for monitoring DEP, the method being A) Diagnosing mood disorders, preferably depressive disorders, in human patients by the method according to the present invention for diagnosing mood disorders, preferably depressive disorders, before initiating treatment that is to be monitored, B) After the period during which the patient received treatment for the depression, steps (a) to (c) of the method according to the present invention for diagnosing a mood disorder, preferably a depressive disorder in a human patient, are repeated to obtain a post-treatment depression score. C) Post-treatment depression score from step (c), - The depression score (i.e., Z-score, probability value) obtained before the patient received treatment for the depression, and - Compared to depression scores (control final values) for normal / healthy subjects, The method includes classifying the treatment as effective if the post-treatment depression score obtained in step B) is closer to the depression score obtained for a normal / healthy subject than the depression score obtained for the period before the patient received treatment.

[0043] In another aspect, the present invention relates to a kit for determining whether a patient is suffering from depression, the kit being 1) Instructions for obtaining an outcome value or depression score by applying the method according to any one of claims 1 to 20, wherein the analysis of the outcome value or depression score determines whether the patient exhibits a mood disorder, preferably a depressive disorder. 2) A set of primers from a group of primer pairs consisting of SEQ ID NOs: 1 to 4 and SEQ ID NOs: 7 to 36, wherein the selection of the set of primers depends on a biomarker used to diagnose depression, or - A combination of at least two sets of primers from the group of primer pairs consisting of Sequence IDs 1 to 36, wherein the selection of the combination of at least two sets of primers depends on a combination of biomarkers used to diagnose mood disorders, preferably depressive disorders. or - The kit includes a set or combination of primers that enables the acquisition of an amplicon, and which includes or is identical to an amplicon obtainable by the set of primers listed in Table 1 (SEQ ID NOs: 1 to 36).

[0044] The following examples, figures, and legend have been selected to provide a complete description to those skilled in the art, enabling them to carry out and use the present invention. These examples are not intended to limit the scope of the invention as the inventors intend, nor are they intended to indicate that only the following experiments were performed.

[0045] Other features and advantages of the invention will become apparent in the remainder of this description, with reference to the examples, figures, and tables set forth below in this specification. [Brief explanation of the drawing]

[0046] [Figure 1A]Identification and validation of A-to-I RNA differential editing sites between healthy controls and DEP patients using RNA-seq data from human blood. A: Genomic location of detected editing sites B: Re-fission of A-to-I editing events or edited genes by chromosome. Dark gray (left y-axis): Re-fission of A-to-I editing events identified with a minimum coverage of 30 in RNA-Seq studies. Light gray (right y-axis): Re-fission of edited genes identified with a minimum coverage of 30 in RNA-Seq studies C: Volcano plot of differentially edited sites between controls and DEP patients. The volcano plot shows upregulatory and downregulatory differentially edited gene sites between the depressed patient (DEP) group and the control group. In each plot, the x-axis represents log(2) (magnitude change) (FC) and the y-axis represents -log10 (p-value). Edited sites with a p-value less than 0.05 and adjusted for FC > 5% are assigned as differentially edited and shown in green. D: Alu editing index between controls and depressed patients. Distribution of Alu editing index (AEI) values ​​in controls and DEP patients. Data are mean ± SEM. The p-value of the editing index was calculated using the Wilcoxon rank-sum test. [Figure 1B] Same as above. [Figure 1C] Same as above. [Figure 1D] Same as above. [Figure 2] Analysis of the main components of 646 variants in 366 differentially edited genes identified in RNA-Seq studies between depressed patients and healthy controls. The scatter plot visualizes the percentage variation of the first, second, and third main components, as well as the ratio (magnification change) of biomarkers, on the x, y, and z axes. Gray circles: DEP patients, black circles: healthy controls. [Figure 3] Protein-protein associations in 14 differentially edited genes used in the priority cohort. The figure shows protein-protein interactions in clusters of 14 genes selected for the priority cohort. Genes are annotated and color-coded regarding their involvement in biological processes.

[0047] [Table 1] [Figure 4A] Examples of diagnostic performance of mRNA editing in the first prioritized cohort (n=76) analyzed by singleplex. A: Example of diagnostic performance obtained by using 4 RNA editing sites on 3 different targets. B: Example of diagnostic performance obtained by using 4 RNA editing sites on 2 different targets. C: Example of diagnostic performance obtained by using 6 RNA editing sites on 4 different targets. D: Example of diagnostic performance obtained by using 29 RNA editing sites present on 5 different targets. [Figure 4B] Same as above. [Figure 4C] Same as above. [Figure 4D] Same as above. [Figure 5A] Diagnostic performance of mRNA editing in a second prioritized cohort (n=86) analyzed by multiplex sequencing. A: Example of diagnostic performance obtained by using two RNA editing sites on three different targets. B: Example of diagnostic performance obtained by using three RNA editing sites on three different targets. C: Example of diagnostic performance obtained by using four RNA editing sites on four different targets. D: Example of diagnostic performance obtained by using five RNA editing sites present on four different targets. E: Example of diagnostic performance obtained by using 19 RNA editing sites on eight different targets. F: Example of diagnostic performance obtained by using 48 RNA editing sites present on nine different targets. [Figure 5B] Same as above. [Figure 5C] Same as above. [Figure 5D] Same as above. [Figure 5E] Same as above. [Figure 5F] Same as above. [Figure 6A]Diagnostic performance of mRNA editing in a validation cohort (n=411) analyzed by multiplex sequencing of eight targets. A: Example of diagnostic performance obtained by using 17 RNA editing sites with five different targets. B: Example of diagnostic performance obtained by using 53 RNA editing sites with eight different targets. C: Example of diagnostic performance obtained by using 79 RNA editing sites with eight different targets. [Figure 6B] Same as above. [Figure 6C] Same as above. [Figure 7A] Diagnostic performance of mRNA editing in a female validation cohort (n=277) analyzed by multiplex sequencing of eight targets. A: Example of diagnostic performance obtained using 11 RNA editing sites on four different targets. B: Example of diagnostic performance obtained using 30 RNA editing sites on six different targets. C: Example of diagnostic performance obtained using 49 RNA editing sites on eight different targets. D: ROC curve obtained after machine learning using the mROC method. E: ROC curve obtained after machine learning using the random forest method. [Figure 7B] Same as above. [Figure 7C] Same as above. [Figure 7D] Same as above. [Figure 7E] Same as above. [Figure 8A] Diagnostic performance of mRNA editing in a male validation cohort (n=134) analyzed by multiplex sequencing of eight targets. A: Example of diagnostic performance obtained by using eight RNA editing sites with six different targets. B: Example of diagnostic performance obtained by using 39 RNA editing sites with seven different targets. C: Example of diagnostic performance obtained by using 68 RNA editing sites with eight different targets. D: ROC curve obtained after machine learning using the mROC method. E: ROC curve obtained after machine learning using the random forest method. [Figure 8B] Same as above. [Figure 8C] Same as above. [Figure 8D] Same as above. [Figure 8E] Same as above. [Figure 9] Correlation between clinical MADRS and IDS-C30 scores for subjects included in RNA-Seq studies. Graphs show 95% confidence intervals. Pearson correlation coefficients and p-values ​​are shown. [Figure 10A] Overview of the Editome Analytics Pipeline [Figure 10B] Same as above. [Figure 11] Demographic characteristics and psychiatric diagnoses of the study population included in RNA-Seq studies. Data are mean ± SEM. p-values ​​for key characteristics are shown using Student's t-test. CTRL: Healthy control; DEP: Major depressive episode; BMI: Body Mass Index. [Figure 12] Functional classification of 366 genes differentially edited between DEP and control based on gene ontology (GO) annotation. Presented are the results of the top 20 GO over-representation tests in the biological process GO term enrichment of the differentially edited genes. Input to the GO enrichment analysis tool was a list of 366 genes differentially edited between DEP and control. The ratio of the number of genes in the GO term to the total number of genes, and p-values ​​corrected using the Benjamini-Hochberg method for multiple testing are presented. FDR, false positive rate; GO, gene ontology. [Figure 13] Signaling pathway enrichment analysis of 366 genes differentially edited between DEP and control. Table 3 shows significant pathways obtained by overpopulation analysis in Reactome. The ratio of the number of genes in a pathway to the total number of genes, and the p-value corrected using the Benjamini-Hochberg method for multiple testing are shown. [Figure 14]Genes and RNA editing variants analyzed in a prioritized cohort. Shown are the gene and location (GRCh38) of the editing variant, post-RNA-Seq coverage, metric change between control and DEP patients, p-values ​​calculated from RNA-Seq data, and area under the curve. p-values ​​were calculated using the Wilcoxon rank-sum test. [Figure 15] Functional classification of 14 differentially edited genes used in a prioritized cohort. Based on gene ontology (GO) annotation. Shown are the results of GO overpopulation tests for the top 20 differentially edited genes in biological process GO term enrichment. Input to the GO enrichment analysis tool was a list of 14 differentially edited genes compared to DEPs and controls analyzed in the prioritized cohort. P-values ​​corrected using the Benjamini-Hochberg method for multiple testing are presented. FDR, false positive rate; GO, gene ontology. [Figure 16] Signaling pathway enrichment analysis of 14 genes analyzed in a prioritized cohort. The table shows significant pathways obtained by overpopulation analysis using Reactome. p-values ​​corrected using the Benjamini-Hochberg method for multiple testing are shown. [Figure 17] Demographic characteristics and psychiatric diagnoses of the study population included in the first priority study analyzed using singleplex analysis. Data are mean ± SEM. P-values ​​for key features are shown using Student's t-test. CTRL: healthy control; DEP: major depressive episode; BMI: body mass index test. [Figure 18] Demographic characteristics and psychiatric diagnoses of the study population included in the second priority study, analyzed by multiplex sequencing. Data are mean ± SEM. P-values ​​for key features are shown using Student's t-test. CTRL: healthy control; DEP: major depressive episode; BMI: body mass index. [Figure 19]Demographic characteristics and psychiatric diagnoses of the study population included in the validation cohort. Data are mean ± SEM. p-values ​​for key characteristics are shown using Student's t-test. CTRL: healthy control; DEP: major depressive episode; BMI: body mass index. [Figure 20] Functional classification of 366 genes differentially edited between DEP and control, based on g:Profiler annotation. Shown are the results of the top 20 GO overpopulation tests in the biological process GO term enrichment with the differentially edited genes. Input to g:Profiler was a list of 366 genes differentially edited between DEP and control. The ratio of the number of genes in the GO term to the total number of genes, and p-values ​​corrected using the Benjamini-Hochberg method for multiple testing are presented. FDR, false positive rate; GO, gene ontology. [Figure 21] Functional classification of 366 genes differentially edited between DEP and control, based on GOnet annotation. Presented are the results of the top 20 GO overpopulation tests in the biological process GO term enrichment with the differentially edited genes. The input to GOnet was a list of 366 genes differentially edited between DEP and control. The ratio of the number of genes in a GO term to the total number of genes, and p-values ​​corrected using the Benjamini-Hochberg method for multiple testing are presented. FDR, false positive rate; GO, gene ontology. [Figure 22A] Diagnostic performance of mRNA editing in a validation cohort (n=411) analyzed by multiplex sequencing of eight targets using random forests. A. Receiver operational characteristic curves and diagnostic performance of depression in the first model adjusted for age, sex, psychiatric treatment, and addiction. The plotted figures represent the probability of a correct response in the tested group. B. Receiver operational characteristic curves and diagnostic performance of depression in the second model adjusted for age and sex. The plotted figures represent the probability of a correct response in the tested group. [Figure 22B] Same as above.

[0048] Example 1: Materials and Method Target population and clinical evaluation Patients with depression (n=268) were recruited from the outpatient department of the Department of Emergency Psychiatry and Post-Acute Care (CHRU of Montpellier) in accordance with the principles of the 1975 Declaration of Helsinki and its subsequent updates. This study was approved by the French local Ethical Committee (CPP Sud-Mediterranee IV in Montpellier, CPP No. A01978-41). All participants aged 18–65 years understood and signed written informed consent before participating in the study. The study was approved by the local institutional review board in accordance with approval requirements and good clinical practice. All patients met the DEP criteria of the Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) using the Mini-International Neuropsychiatric Interview. The presence or absence of psychiatric diagnoses and mood states at the time of participation in the study was confirmed by a trained psychiatrist. During standardized interviews, psychiatrists administered the French version of the Montgomery-Asberg Depression Rating Scale (MADRS) and the Clinician Assessment (IDS-C30), a 30-item Inventory of Depressive Symptomatology, to score depression. All white DEP patients were recruited from the outpatient department of the Department of Emergency Psychiatry Post-Acute Care (CHRU of Montpellier). Age, race, and sex-matched controls (n=143) were recruited from a list of volunteers from the Clinical Investigation Center (CHRU of Montpellier).

[0049] RNA extraction and qualification screening from whole blood. Four ml volumes of whole blood from each patient were collected in PAXgene® blood RNA tubes optimized for RNA stabilization, stored at -20°C, and then transferred to -80°C. The samples were randomly dispersed into different sets of extracts. Blood samples were thawed overnight at 4°C and centrifuged at 3000 g for 10 minutes. The pellets were washed with 4 ml of 1x DPBS and resuspended in 400 μl of 1x DPBS. Total RNA from the samples was isolated using the MagNA Pure 96 Cellular RNA Large Volume Kit (LifeScience) according to the manufacturer's automated protocol. RNA was eluted in 100 μl of MagNA Pure 96 elution buffer. Standard precautions were taken to avoid RNA degradation by RNA-degrading enzymes (RNAse) during sample preparation and RNA extraction. Total RNA concentration and quality levels were measured using a Qubit Fluorometer (Invitrogen) and a LabChip (Perkin-Elmer, HT RNA Reagent Kit) instrument, respectively.

[0050] Library preparation for RNA sequencing To prepare an RNA library from blood-derived RNA, a specific strategy was selected to enrich the library with total RNA (mRNA) by depleting both ribosomal RNA and globin RNA, the two abundant forms of RNA, in a single step. Total RNA was preferred over messenger RNA because RNA editing in hematopsies has been shown to enrich the library with intron regions of repeat elements that are removed during splicing to generate messenger RNA. Therefore, a TruSeq Stranded Total RNA Library Kit (Illumina) containing Ribo-Zero Globin, specially formulated for blood samples, was used according to the manufacturer's instructions. Briefly, 300 ng of total RNA was depleted in ribosomal and globin RNA using rRNA and globin depletion probes combined with magnetic beads. Following purification, the RNA was fragmented into averaging 250 bp fragments using divalent cations at high temperature. First strand cDNA synthesis was then achieved using Superscript II reverse transcriptase (Thermo Fisher Scientific) and random primers. Next, second-strand cDNA synthesis was performed using DNA polymerase I and RNase H. Importantly, during second-strand cDNA synthesis, dUTP was incorporated instead of dTTP to generate a strand-specific library. Following second-strand cDNA synthesis, and prior to adapter ligation, the 3' ends of the blunt-end fragments were 3' adenylated. Importantly, each index was associated with all groups to prevent inference bias due to the non-specific effect of the index on sequencing depth. The DNA fragments with adapter molecules at both ends were then amplified by PCR to increase the amount of DNA in the library. Quality control was then performed using both Qubit Fluorometer (Invitrogen) and LabChip (Perkin-Elmer, HT DNA 5K Reagent Kit) instruments for quantification and quality control analysis of the DNA library template, respectively.Next, the samples were pooled into a library and purified using the Magbio PCR Cleanup System. The library was denatured using 0.1 M NaOH and sequenced on an Illumina NextSeq 500 / 550 High Output using 75 bp read length paired-end chemistry.

[0051] Processing of RNA-Seq data, differential expression analysis, and editome analysis Paired-end reads were generated using an Illumina NextSeq 500 and demultiplexed using bcl2fastq (version 2; 17.1.14, Illumina). Sequencing quality was performed using FastQC software (version 0.11.7, https: / / github.com / s-andrews / FastQC). The STAR aligner 6 was used to perform read alignment using the human genome version hg38, followed by realignment, removal of PCR duplicates, and final single nucleotide variant (SNV) calling by the Genome Analysis Tool Kit (GATK version 4.1.4.0) 7 RNAEditor (version 1.0) 8 was used to identify and quantify A-to-I editing events. Then, ANNOVAR 9 , RepeatMasker 10 , and BiomaRt 11 were used to perform further filtering and functional annotation of the edited events.

[0052] Preparation and sequencing of targeted next-generation sequencing libraries For the preparation of the next-generation sequencing (NGS) library, a multiplex targeting approach was chosen to selectively sequence the target region within each relevant target. The target region was amplified by PCR using validated PCR primers (Table 20). For PCR amplification, the Q5 Hot Start High Fidelity enzyme (New England Biolabs) was used according to the manufacturer's guidelines. PCR responses were performed on a Peqstar 96x thermocycler using an optimized PCR protocol. Both the quantity and quality of the PCR products were evaluated. The purity of the amplicons was determined using a Nucleic Acid Analyzer (LabChipGx, Perkin Elmer), and quantification was performed using the fluorescence-based Qubit method. After quality control, the PCR response was purified using magnetic beads (High Prep PCR MAGbio system, Mokascience). DNA was then quantified using the Qubit system, and the purification yield was calculated. Next, samples were individually indexed by PCR amplification using the Q5 Hot-start High-fidelity PCR enzyme and the Illumina 96 Indexes kit (Nextera XT Index Kit; Illumina). The samples were then pooled into a library and purified using the Magbio PCR Cleanup System. The library was denatured with 0.1M NaOH and loaded into sequencing cartridges (Illumina MiSeq Reagent Kit V3 or Illumina NextSeq 500 / 550 Mid-Output) according to Illumina guidelines. A commercially available total RNA pool from human blood peripheral leukocytes (Clontech, #636592) was incorporated into the library to determine variability between different sequencing flow cells during the experiment. The NGS library was spiked to introduce library diversity using PhiX Control V3 (Illumina), and sequencing was performed at standard concentrations using deep sequencing (>50K sequences per sample) (single-read sequencing, read length 150 bp).

[0053] Bioinformatics analysis of targeted sequencing data Sequencing data was downloaded from a NextSeq or MiSeq sequencer (Illumina). Sequencing quality was performed using FastQC software (version 0.11.7, https: / / github.com / s-andrews / FastQC / ). For further analysis, the minimum sequencing depth of 20,000 reads for each sample was considered. A preprocessing step was performed, consisting of removing adapter sequences and filtering sequences according to size and quality score. Short reads (<100 nts) and reads with an average QC <20 were removed. To improve the sequence alignment quality of the sequences, a flexible read trimming and filtering tool for Illumina NGS data was used (fastx_toolkit v0.0.14 and prinseq version 0.20.4). After the preprocessing step, additional quality control was performed on each cleaned fastq file before further analysis.

[0054] bowtie2 in end-to-end sensitive mode 12 Alignment of processed reads was performed using version 2.2.9 of the SAMtools software. The alignment was performed against the most recent reference human genome sequence (GRCh38). 13 Version 1.7 removed non-specific alignments, reads, unaligned reads, or reads containing insertions / deletions (INDELs) from downstream analysis. SVN calling now uses SAMtools mpileup. 14We used the following: We studied edited sites within the alignment by using an in-house script to count the number of different nucleotides at each genomic location. We calculated the percentage of reads carrying nucleotide "G" instead of nucleotide "A" at each location [number of "G" reads / (number of "G" reads + number of "A" reads) * 100]. Edit rates >0.1 were automatically detected by the script and considered "A-to-I edit sites". The final step was to calculate the percentage of all possible isoforms for each transcript. By definition, the relative percentage of RNA edits at a given edit "site" represents the total number of edit modifications measured at this unique genomic coordinate. Conversely, an mRNA isoform is a unique molecule that may or may not contain multiple edit modifications in the same transcript. For example, in a given transcript, mRNA isoform BC contains A-to-I modifications at both sites B and C within the same transcript. We set a threshold of at least 0.1% relative percentage for inclusion in the analysis.

[0055] Statistical analysis of data All statistics and figures are from the "R / Bioconductor" statistics open-source software. 15 Alternatively, calculations were performed using GraphPad Prism software (version 7.0). Biomarker values ​​(i.e., RNA editing sites and isoforms of target genes, as well as mRNA expression of ADARs) are typically presented as mean ± mean standard error (SEM). To ensure normal distribution of data, all biomarkers may be converted using the bestNormalize R package (version 1.4.2). Differential analysis was performed using the most appropriate test among the Mann-Whitney rank-sum test, Student's t-test, or Welch's t-test, according to normality and sample variance distribution. A p-value less than 0.05 was considered statistically significant.

[0056] The diagnostic performance of a biomarker can be characterized by susceptibility, which represents the ability to detect the “depression” group, and specificity, which represents the ability to detect the “control” group. The evaluation results of the diagnostic test can be summarized in a 2x2 contingency table comparing these two clearly defined groups. By setting a cutoff, the two groups can be categorized according to the test results, either positive or negative. Considering a particular biomarker, it is possible to identify the number of patients A (“true positive”: TP) with positive test results among the depression group, and patients B (“true negative”: TN) with negative test results among the “control” group. In the same manner, patients C (“false negative”: FN) with negative test results among the “depression” group, and patients D (“false positive”: FP) with positive test results among the “control” group are observed. Susceptibility is defined as TP / (TP+FN), which is referred to herein as the “true positive rate.” Specificity is defined as TN / (TN+FP), which is referred to herein as the “true negative rate.”

[0057] Receiver Operating Characteristics (ROC) Analysis 16 The precision and distinguishing power of each biomarker were evaluated using the ROC curve. The ROC curve graphically visualizes the interrelationship between sensitivity (Se) and specificity (Sp) of various test values. In addition, all biomarkers were combined with each other, for example, in the mROC program. 17 Or Random Forest 18 We used multiple approaches to assess the potential increase in sensitivity and specificity.

[0058] mROC is a dedicated program for identifying linear combinations and maximizing AUC (Area Under Curve) ROC. Equations for each combination are provided and can be used as the following new virtual marker Z: Z = a × biomarker 1 + b × biomarker 2 + c × biomarker 3 In the formula, a, b, and c are the calculated coefficients, and biomarkers 1, 2, and 3 are the levels of the biomarkers considered.

[0059] The aforementioned Random Forest (RF) was applied to evaluate RNA editing isoform / site combinations. This method combines the concept and features of random selection from Breiman's "bagging" to construct a collection of variance-controlled decision trees. Using Random Forest, the importance of isoform / site editing can be ranked, and the best isoforms can be combined. Multiple Downsizing (MultDS) asymmetric bagging with embedded variable classifiers was performed. This method attempts to utilize majority calls of all information through multiple downsampling while keeping the minority classes fixed. The algorithm generated 100 Random Forests from the training set, each containing the same fixed number of minority classes and an equal or adjusted number of randomly sampled elements, producing a nearly balanced tree. For classification of new data, the subsequent results were joined by majority vote of the 100 trained Random Forests. Tenfold cross-validation was applied to estimate the RF parameters for each of the 100 trained Random Forests. The implementation was performed using the R Random Forest package (version 4.6-14) and the R caret package (version 6.0-84).

[0060] Example 2: Results Characteristics of Discovery Study Samples: Patients and controls (n=32 in each condition) included in the RNA-Seq samples were matched for all variables that did not differ statistically between groups, such as sex, age, or BMI (Figure 11). To confirm the presence or absence of a psychiatric diagnosis, both groups selected two different clinical assessments designed by the American Psychiatry Association Diagnostic and Statistical Manual of Mental Disorders: MADRS and IDSC-30. Healthy controls did not show signs of depressed mood at the time of sample collection (MADRS<7 and IDSC-30<10). Patients with depression (DEP) had significantly higher scores on both MADRS (p-value 5.93E-09) and IDSC-30 (p-value 2.59E-09) (Table 1). Furthermore, a significant association (r²=0.84, p<0.0001) was observed between the two clinical assessments (Figure 9), suggesting correct patient assignment and relative homogeneity in depression scoring.

[0061] Editome analysis To obtain an overall picture of RNA editing event modifications in depressed patients, we analyzed RNA-Seq datasets using our in-house editome analysis pipeline (see Methods and Figure 10). To ensure high reliability at specific base locations, we identified 37,599 A-to-I differential edited sites with a minimum coverage of 30x. Coverage describes the average number of reads that align to, or "cover," a known reference base. The majority of edited sites (32,183 sites, 85.6% of the total) were located in introns (Figure 1A), consistent with other studies in human tissue. Consistent with previous reports using RNA-Seq, the percentage of edited sites in the 3' untranslated region (3'UTR) of mRNA was 6.2% (2344 sites). The number of edited sites was uniformly redistributed across the genome, although edited genes on chromosomes 4, 9, 18, 19, 20, 21, and 22 were slightly excessive (Figure 1B). Next, differential analysis was performed to identify sites where editing might differ specifically between DEPs and healthy controls. After applying pre-specified quality criteria (exclusion of inter-gene sites, coverage > 30, AUC > 0.6, 0.95 > multiplier change > 1.05, and p < 0.05), 646 variants representing 366 genes that were differentially edited between depressed patients and healthy controls were identified (Figure 1C). The quality criteria were set to provide potential diagnostic biomarkers (BMKs) for depressive disorders with clinical utility. Importantly, no differences in global RNA editing were observed between patients and controls. In fact, the global Alu editing index (AEI), a measure of the overall proportion of RNA editing in Alu repeats, was calculated. Alu repeats are the most common edited regions in the human genome and are therefore an indicator of the overall editing rate. No significant difference in AEI was found between controls and depressed patients (AEI control mean = 0.300 ± 0.002, AEI DEP mean = 0.299 ± 0.003, p-value = 0.57, Figure 1D). Since the Alu editing index characterizes the weighted average editing level across all expressed Alu sequences, these data suggest that differential RNA editing is target-specific rather than disease-dependent.

[0062] Gene ontology tools 19 GSEA (Genetic Set Enrichment Analysis) of these 366 genes using G:profiler showed strong term enrichment (false detection rate <0.05) for biological processes including multiple immune categories (response, activation, and regulation), cellular activation, and metabolic processes (Figure 12). This analysis was performed using G:profiler 20 (Figure 9) or GOnet 21 These findings were reproduced using other tools, such as (Figure 10), and consistently demonstrated robust enrichment of biological processes related to the immune system. Furthermore, reaction pathway analysis of these 366 genes revealed robust enrichment in immune system-related pathways, namely TCR signaling, ZAP-70 translocation to immune synapses, or interferon-gamma signaling (Figure 13). Notably, principal component analysis (PCA) showed that this subset of 646 edit variants representing the 366 genes clearly distinguished between two distinct groups: controls and depressed patients (Figure 2), suggesting that these genes can be combined with edit variants to selectively differentiate depressed patients from healthy controls.

[0063] To further refine the list of genes, we used much stricter quality criteria (coverage > 30, AUC > 0.8, 0.8 > 50% change > 1.20, and p < 0.05). We then manually curated databases of these markers to arrive at several combinations of biomarkers representing different biological mechanisms. Using these secondary criteria, we selected a list of 15 variants representing 13 genes (Figure 14) for both diagnostic performance and their potential roles in the immune system or psychiatric biology, and studied them in subsequent prioritized cohorts. In addition, previous data showed edit changes in the brains of suicide victims and in the blood of drug-induced depression in HCV patients. 5、22 , and specified PDE8A as the other target. STRING database 23Protein-protein network analysis using [tool name] revealed a network of enriched proteins in protein-protein interactions (PPI p-value < 4.5.10-6). Interactions within the STRING are evidence-based and consist of direct (physical) and indirect (functional) interactions. This specification focuses on the minimum interaction score of 0.7 required, which is an indicator of high reliability in actual protein-protein interactions. Furthermore, gene ontology tools were used. 19 Functional protein-associated network analysis of 14 selected genes using this method showed robust enrichment of genes associated with immune system processes (8 genes, GO:0002682, p(FDR=4.10-4), regulation of immune responses (6 genes, (GO:0050776, p(FDR=2.2.10-3)), or typically, the JAK-STAT-mediated receptor signaling pathway triggered after interferon-alpha stimulation (4 genes, GO:0007259) (Figures 3 and 15). Furthermore, Reactome pathway analysis showed that these 14 genes were associated with immune system-related pathways (e.g., IFNA signaling (HSA-912694, p(FDR)<0.025) or cytokine signaling (HSA- It was shown to be enriched in the regulation of 1280215 (p(FDR)<0.026), and neurotransmitter regulation (e.g., AMPA receptor transport (HSA-399719, p(FDR)<0.025), glutamate binding, and synaptic plasticity (HAS-399721, p(FDR)<0.025), or neurotransmitter receptor and postsynaptic signaling (HSA-112314, p(FDR)<0.029)) (Figure 16). Interestingly, the top-level pathway identified by Reactome analysis (p(FDR)<0.012) is linked to the transcription factor Runx1, which has recently been shown to mediate the effects of chronic social defeat stress in a mouse depression model through the regulation of miRNAs of the miR-30 family.

[0064] Target description PRKCB: Protein kinase C-beta (PRKCB), belonging to the family of serine and threonine-specific protein kinases, is known to play an important role in the immune system, as PRKCB knockout mice develop immunodeficiency characterized by B cell activation, impaired B cell receptor response, and defects in T cell-independent immune response. Furthermore, PRKCB acts as a regulator of the HPA axis response to stress. 24 A study of 44 candidate genes associated with depression and anxiety symptoms in postpartum women revealed a link between PRKCB and major depression. In fact, PRKCB interacts with CREB1, which directly modulates the expression of several genes involved in the brain's stress response, such as BDNF, the BDNF receptor Trk-β, and glucocorticoid receptors, and phosphorylates CREB1. 25 .

[0065] MDM2: MDM2 is a major cell inhibitor of p53 and can be a target for anti-cancer therapy. However, MDM2 also catalyzes the ubiquitination of β-arrestin, a target of antidepressants. In the CNS, MDM2 is involved in AMPAR (glutamate receptor) surface expression during synaptic plasticity. 26 Interestingly, RNA editing in MDM2 has been shown to regulate protein levels. In fact, increased RNA editing at the 3'UTR of MDM2 may negate microRNA-mediated repression and increase mRNA levels.

[0066] PIAS1: PIAS1 (an activated STAT1 protein inhibitor) functions as a SUMO ligase that blocks the DNA-binding activity of Stat1 and inhibits Stat1-mediated gene activation in response to interferon. SUMOylation of the transcription factor (HSA-3232118) is one of the major Reactome pathways of 14 genes selected for analysis in a prioritized cohort (Figure 16). PIAS1 selectively inhibits interferon-induced genes and is a key player in innate immunity. 27In addition, Pias1 interacts with presynaptic metabotropic glutamate receptor 8 (mGluR8), initiating post-transcriptional modification via SUMOylation, suggesting that PIAS1 may have a regulatory function in targeting and regulating mGluR8 receptor signaling.

[0067] IFNAR1: Numerous studies have linked interferon to inflammation-induced changes in brain function and depression, at least partially through the effects of IL-6, CXCL10, or by the induction of indoleamine 2,3-dioxygenase 1 (IDO1). 28 Furthermore, polymorphisms in the promoter region of IFN-alpha / beta receptor 1 (IFNAR1) have been shown to influence the risk of developing depression. Interferons are used to treat viral hepatitis, hematopoietic disorders, autoimmune disorders, and malignancies. However, regardless of the outlook for IFN therapy, IFN is associated with a 30–70% risk of treatment-induced depression. Previous studies have shown that RNA editing biomarkers have made it possible to distinguish between patients who developed treatment-induced depression and those who did not.

[0068] IFNAR2: Interferons act through receptors, and the interferon type I receptor (IFNaR) is composed of two subunits, IFNAR1 and IFNAR2. Polymorphisms in the IFNAR2 gene (SNP11876) were associated with susceptibility to multiple sclerosis (p=0.035). Interestingly, the same study also identified polymorphisms in the IFNAR1 gene (SNP18417), another member of the same receptor.

[0069] LYN: LYN encodes a non-receptor tyrosine protein kinase that transmits signals from cell surface receptors. LYN plays a crucial role in regulating innate and adaptive immune responses. 29In the mammalian central nervous system, Lyn is highly expressed in the nucleus accumbens, cerebral cortex, and thalamus. Furthermore, Lyn is involved in the regulation of the N-methyl-D-aspartate receptor (NMDAR, ion channel glutamate receptor) pathway. In fact, behavioral studies have shown that spontaneous motor activity is suppressed in Lyn- / - mice due to enrichment of NMDA receptor signaling. In addition, Lyn physically interacts with AMPA receptors, which are members of the glutamate receptor family. 30 Following receptor stimulation, Lyn activates the mitogen-activated protein kinase (MAPK) signaling pathway, thereby increasing the expression of brain-derived neurotrophic factor (BDNF). Importantly, the activation of AMPA receptors and downstream synaptic signaling pathways associated with BDNF release mediates the antidepressant effects of ketamine.79 The Lyn kinase-glutamate receptor interaction is dramatically adapted in mood disorders such as major depressive disorder.

[0070] AHR: Aryl hydrocarbon receptors (AHRs) are highly conserved ligand-dependent transcription factors that sense environmental toxins and endogenous ligands. AHRs also regulate the differentiation and response of immune cells. 31 AHR also regulates the differentiation of both T(reg) and Th17 cells in a ligand-specific manner. Activated AHR induces IDO and therefore increases kynurenine levels. The role of kynurenine in the CNS, which plays a crucial role in the pathophysiology of pro-inflammatory depression, is well documented. Pharmacological blockade of AHR rescued systemic pro-inflammatory monocyte transport, neuroimmunodeficiency, and depression-like behavior in mice. RNA editing affects the 3'UTR of human AHR mRNA, thereby creating a microRNA recognition sequence that modulates AHR expression.

[0071] RASSF1: Ras-related domain family member 1 (RASSF1) RASSF1A is a tumor suppressor gene whose inactivation is associated with a wide variety of sporadic human cancers and is frequently epigenetically inactivated in various solid tumors. RASSF1A associates with MDM2 and enhances the auto-ubiquitin ligase activity of MDM2. 32 Interestingly, microarray analysis has recently identified RASSF1A as a gene that is significantly downregulated in women with a history of postpartum depression.

[0072] GAB2: The Gab protein functions downstream of multiple receptors, including those for EGF, hepatocyte growth factor, and FGF. GAB2 is a key component of neuronal differentiation induced by retinoic acid, and partially by acting downstream of bFGF, mediating survival via PI3 kinase-AKT. GAB2 positively upregulates the expression of IL-4 and IL-13 in activated T cells, suggesting that GAB2 may be a key regulator of the human Th2 immune response. 33 .

[0073] CAMK1D: CAMK1D is a member of the calcium / calmodulin-dependent protein kinase 1 family and is involved in chemokine signaling pathways that regulate granulocyte function. 34 In the CNS, CAMK1D is primarily expressed in the CA1 region of the hippocampus and is induced following long-term strengthening (LTP), which is a prolonged enhancement of synaptic effects in the central nervous system. It can translocate to the nucleus in response to stimuli that cause intracellular Ca2+ influx, such as glutamate, and activate CREB-dependent gene transcription. Notably, the CAMK1D SNP rs2724812 has been associated with remission after selective serotonin reuptake inhibitor therapy using a deep learning approach. 35 .

[0074] KCNJ15: Potassium voltage-gated channel subfamily J member 15 (KCNJ15) encodes the inward-rectifying potassium channel Kir4.2. In the CNS, the KIR4.2 channel is expressed in the later stages of development in both the mouse and human brains. Inward-rectifying potassium channels (Kir channels) are important regulators of neuronal signaling and membrane excitability. Gating of Kir channel subunits plays a crucial role in polygenic CNS diseases. Furthermore, rs928771 of KCNJ15 has been found to be associated with alterations in whole blood KCNJ15 transcription levels, and network analysis of suggested hippocampal and blood transcriptome datasets suggests that risk variants at the KCNJ15 locus may act through their regulatory effects on immune-related pathways. 36 .

[0075] IL17RA: Interleukin-17 receptor A (IL17RA) is a type I membrane glycoprotein with a structure similar to IFNAR1, and binds to interleukin-17A with low affinity. Interleukin-17A (IL17A) is a pro-inflammatory cytokine primarily secreted by activated Th17 lymphocytes. Upon contact with its congener receptor, IL17 transmits signals through the ERK, MAPK, and NFKB pathways, all of which are involved in depressive disorders. Th17 cells target a variety of brain commensal cells, including neurons, astrocytes, and microglia, and are involved in neuroinflammation, making them a potent regulator of brain function. 37IL-17RA is highly expressed in blood-brain barrier (BBB) ​​endothelial cells of multiple sclerosis lesions, and IL-17 increases BBB permeability by disrupting tight junctions, thereby promoting CNS inflammation. Pro-inflammatory IL-17 is also secreted by activated astrocytes and microglia and has detrimental effects on neurons. Flow cytometry analysis of PBMCs in unmedicated depressed subjects revealed a significant increase in peripheral Th17 cell count and elevated serum IL-17 concentrations, suggesting a potential role of Th17 cells in MDD patients. Similarly, Th17 cells increase in the brains of mice after two common depression-like models: learned helplessness or chronic restraint stress. Inhibition of Th17 cell production or function reduces vulnerability to depression-like behavior in the same mouse models.

[0076] PTPRC, also known as the leukocyte common antigen or CD45, is a protein tyrosine phosphatase, type C receptor (PTPRC), which contains an external segment and is capable of initiating transmembrane signaling in response to an external ligand. 38 In the immune system, PTPRCs are essential for the activation of T and B cells by mediating cell-cell contact and regulating protein-tyrosine kinases involved in signal transduction. PTPRCs repress and negatively regulate cytokine receptor signaling, and selective disruption of PTPRC expression results in enhanced interferon receptor-mediated activation of JAK and STAT kinase proteins. 39Notably, point mutations in PTPRC (a C-to-G nucleotide transition at position 77 of exon 4) are associated with the development of multiple sclerosis, the most common demyelinating disease of the CNS. In the brain, anxiety-like behavior induced by repeated social defeats in a mouse model of depression corresponded to an increase in brain macrophages overexpressing PTPRC. In the same mouse model of depression, classical benzodiazepines were shown to reverse symptoms and block stress-induced accumulation of macrophages expressing high levels of PTPRC in the CNS. Furthermore, in patients diagnosed with major depressive disorder, PTPRC levels were higher in patients than in healthy subjects. The PTPRC ratio decreased with antidepressant treatment such as venlafaxine or fluoxetine and was positively correlated with the severity of depression. 40 This suggests that antidepressant treatment contributes to immunomodulation in patients with major depressive disorder.

[0077] PDE8A: The glutamate or serotonin receptor is a G protein-coupled receptor (GPCR) involved in a variety of mental disorders, where spatial and temporal regulation of second messenger concentrations is crucial for subsequent signaling. Phosphodiesterases (PDEs) are important modulators of signaling and are involved in inflammatory cell activation, behavior, memory, and cognition. PDE8A is a cAMP-specific phosphodiesterase expressed in the brain and blood. Alterations in RNA editing of the PDE8A transcript have been identified in T cells from patients with systemic lupus erythematosus, a chronic autoimmune disorder, as well as in normal T cells with IFN-α. Phosphodiesterase 8A (PDE8A) expression is reduced by half in the temporal cortex of patients with major depressive disorder (MDD). 41 In recent years, brain region-specific alterations in RNA editing of PDE8A mRNA in the offspring of suicide victims have been demonstrated, providing the first immune response-related brain marker in suicide victims. 5PDE8A RNA editing is also increased in SHSY-5Y neuroblastoma cells treated with interferon-alpha, as well as in HCV patients receiving antiviral combination therapy with IFN-α and ribavirin. Importantly, measuring RNA editing in PDE8A allows for differentiation between patients who developed treatment-induced depression and those who did not, suggesting that AI RNA editing biomarkers may be useful for monitoring treatment-induced depression. 22 .

[0078] To prioritize a subset of clinically useful variants for diagnostic assays, each of the target genes listed above was individually tested in two separate “prioritization” cohorts (n=76, Figure 17 and n=86, Figure 18) by measuring the edit rates of controls and depressed patients using NGS. In the first cohort, each target was sequenced individually, while in the second cohort, the edit rate of biomarkers was measured by a multiplex of four BMKs. Multivariate statistical analysis methods were performed to evaluate each combination of RNA-edited sites and / or isoforms to find the best combination of biomarkers. To enhance diagnostic performance while minimizing the number of biomarkers, the best combinations using different numbers of biomarkers were calculated.

[0079] Table 1 shows a list of primers used to detect RNA editing sites.

[0080] [Table 2-1]

[0081] [Table 2-2]

[0082] Since it was important to validate the results obtained from RNA-Seq studies, we first performed differential analysis of each target in a prioritized cohort. As shown in the following tables (Tables 2A-2B and 3), all sites and isoforms identified in RNA-Seq studies were found to be significant in subsequent analyses in larger populations.

[0083] Tables 2A and 2B: Differential analysis of edit sites (2A) and isoforms (2B) identified in RNA-Seq studies of patients in the first priority cohort analyzed using singleplex analysis.

[0084] Shown are examples of RNA-Seq data for different targets, including target name, differentially edited site or isoform, p-value and magnification change of the edited site or isoform, AUC ROC, mean value of edits at the considered site, and population size.

[0085] [Table 3]

[0086] [Table 4]

[0087] Table 3: Differential analysis of edit sites and isoforms identified in RNA-Seq studies of patients in the second priority cohort analyzed by multiplex sequencing.

[0088] Shown are examples of RNA-Seq data for different targets, including target name, differentially edited site, p-value and magnification change of the edited site, AUC ROC, mean value of edits at the considered site, and population size.

[0089] [Table 5]

[0090] Next, biomarker combinations were analyzed to identify the best diagnostic performance with different numbers of biomarkers. In the first cohort, using only four sites from three different targets yielded an AUC of 0.811 (Sp 78.1%, Se 81.3%, Figure 4A). Another combination of four sites using two BMKs yielded even better performance with an AUC of 0.866 (Sp 92.7%, Se 79.0%, Figure 4B). The equation used for this combination is as follows: Z=+(0,361161202728045)*IFNAR1_X227_H+(-0,0763586204473731)*IFNAR1_ X162_E+(1,35409123139041)*LYN_X284_P+(-2,2905526156995)*LYN_X207_M

[0091] In the measurement of RNA editing variants, good performance was obtained for all combinations tested, but the performance did not necessarily reflect the number of biomarkers. When the number of biomarkers was increased to six, the AUC for four different targets was 0.794 (Sp 80.5%, Se 77.1%, Figure 4C). However, when five BMKs were combined and all differentially edited sites of these targets were analyzed, excellent performance was obtained with an AUC of 0.977, showing superior specificity and sensitivity (92.7% and 94.7%, respectively, Figure 4D). The equation used for this combination is as follows: Z=+(1,242350672584)*IFNAR1_X227_H+(1,36013909308819)*IFNAR1_X115_G+(-0,739944712498771)*IFNAR1_X84_F+(1,42700898962888)*IFNAR1_X162_E+(-1,66838875199491)*IFNAR1_X231_D+(-0,887010245248778)*IFNAR1_X129_B+(0,0876014897108424)*IFNAR1_X111_A+(0,985129912485808)*LYN_X153_G+(4,33197064304577)*LYN_X284_P+(-0,200203331256123)*IFNAR1_X152_C+(-15,708397165937)*IL17RA_X78_D+(2,50160465829557)*LYN_X116_T+(5,53594778946524)*IL17RA_X77_C+(-3,04072407838489)*LYN_X47_Q+(1,82611854035731)*IL17RA_X184_H+(1,14130391422593)*LYN_X85_R+(-1,36538923404951)*LYN_X63_C+(-1,52552008654041)*LYN_X172_J+(3,47820883385469)*LYN_X203_L+(-1,80085106045942)*IL17RA_X55_B+(-1,25213023530841)*LYN_X163_Z+(-1,01733883175651)*PTPRC_X240_I+(122,471673425459)*IL17RA_X151_G+(1,09079061375429)*LYN_X64_D+(2,32804478266105)*LYN_X255_N+(-3,86751458009041)*LYN_X167_I+(-12,6254508356666)*LYN_X148_A+(3,00274639389995)*PTPRC_X299_O-(1,08899988893022)*PDE8A_SiteB

[0092] Table 4 lists all combinations tested in this cohort. It should be noted that in the first cohort, some samples were not analyzed, and the total population of combinations may be smaller than the total cohort population. All combinations have a p-value < 10. -4 The results were statistically significant between the control group and the cases.

[0093] [Table 6-1]

[0094] [Table 6-2]

[0095] [Table 6-3]

[0096] [Table 6-4]

[0097] [Table 6-5]

[0098] [Table 6-6]

[0099] In a second cohort using multiplex sequencing, the AUC was 0.786 when only two sites were selected for each BMK (Sp 78.6%, Se 81.0%, Figure 5A). In this cohort, increasing the number of biomarkers systematically increased the overall diagnostic performance of the assay. Combining three sites from three BMKs increased the AUC to 0.839 (Sp 89.3%, Se 74.1%, Figure 5B). Using four and five sites resulted in specificity of 82.1% and 92.9%, and sensitivity of 81% and 77.6%, respectively, with AUCs increasing to 0.852 and 0.860 (Figures 5C and D). Increasing the combination to include 19 sites resulted in an AUC greater than 0.9 (AUC 0.909, Sp 92.9%, Se 86.2%, Figure 5E). Furthermore, combining 48 differentially edited sites present in nine different BMKs yielded an exceptional AUC of 0.993 with 100% specificity and 96.6% sensitivity (Figure 5F), and the equation used for this combination is as follows: Z=-(0.179462180023875)*MDM2_Site207_W-(-1.04404354492012)*IFNAR1_Site101_nan-(2.64479093229565)*GAB2_Site88-(4.11125599566128)*v1MDM2_Site76-(1.61229526347695)*MDM2_Site143-(5.57184346092994)*LYN_Site90_S-(5.25866173918605)*PDE8A_Site116248_B-(18.995426153835)*KCNJ15_Site134_D-(-1.04699821617189)*v1MDM2_Site86-(0.480007114248953)*CAMK1D_Site131-(3.56662634744781)*KCNJ15_Site130_A-(2.4602823172239)*KCNJ15_Site129_B-(-9.43134276848121)*KCNJ15_Site133_C-(1.42715048538683)*IFNAR1_Site76_nan-(1.00606326866477)*CAMK1D_Site49-(0.0542132983688757)*v1MDM2_Site87_E+(5.91151340839856)*LYN_Site155_X-(-2.20437543623734)*KCNJ15_Site131-(1.30800292033623)*PTPRC_Site169_E-(4.82052635778379)*MDM2_Site277-(-10.1173014238565)*KCNJ15_Site190-(-3.7454636873821)*PTPRC_Site170_F-(1.03362539706716)*PTPRC_Site259-(-4.87935759821363)*PTPRC_Site163_D-(0.628308423459796)*PRKCB_Site119_E-(-0.901688238257015)*KCNJ15_Site193_E-(-0.425246156880022)*MDM2_Site264-(-0.533167398964742)*IFNAR1_Site123_nan-(-5.64307592868436)*IFNAR1_Site131_nan-(47.981969789844)*GAB2_Site131_H-(0.0402950393688472)* MDM2_Site237_X-(-3.33278344933062)*MDM2_Site192-(0.928449146700883)*CAMK1D_Site64_D-(-1.0 8069775385396)*GAB2_Site126_A-(-4.23184777312141)*CAMK1D_Site113_G-(1.99562009099246)*GAB 2_Site52_B-(2.3505776086063)*IFNAR1_Site142_nan-(3.97251790792544)*PRKCB_Site103_C-(0.570 977372558063)*PRKCB_Site126-(-0.410508492870566)*PRKCB_Site132_H-(0.975605316457069)*IFNA R1_Site128_nan-(-1.09288061572838)*PDE8A_Site116278_C+(12.6078350770629)*CAMK1D_Site125-( 1.48817555399057)*PTPRC_Site253-(-0.716028094006562)*PRKCB_Site99-(-0.722585574440859)*IF NAR1_Site151_nan-(0.00161227142613181)*IFNAR1_Site152_C-(-7.23169188321118)*PRKCB_Site183.

[0100] Table 5 presents a list of all combinations tested in the second priority cohort, which included all 86 patients. All combinations showed a p-value < 10. -4 The results were statistically significant between the control group and the cases.

[0101] [Table 7-1]

[0102] [Table 7-2]

[0103] [Table 7-3]

[0104] [Table 7-4]

[0105] [Table 7-5]

[0106] [Table 7-6]

[0107] Overall, these data suggest that DEP can be diagnosed by sequencing using different biomarkers from a panel of 14 genes selected in a prioritized cohort.

[0108] Importantly, all identified biomarkers included in this panel of 14 genes are present in at least one combination, suggesting that DEP can be diagnosed with favorable diagnostic performance using all of these BMKs.

[0109] We selected a panel of eight of the most interesting biomarkers and sequenced them in a multiplex format to further validate the results in a large cohort (validation cohort) of 411 subjects, 143 of whom were healthy controls and 268 were DEP patients (Figure 19).

[0110] Using the machine learning approach Random Forest (RF), we combined all BMKs with each other to assess the potential increase in susceptibility and specificity. Analysis of RNA editing variants of eight BMKs (GAB2, IFNAR1, KCNJ15, LYN, MDM2, PRKCB, CAMK1D, and PDE8A) yielded an AUC ROC curve of 0.929 (CI95%:[0.876-0.981]) with a susceptibility of 82.9 and specificity of 87.1% (Figure 22A) in a Model 1 trial dataset adjusted for age, sex, psychiatric treatment, and addiction.

[0111] [Table 8]

[0112] In addition, a random forest model 2, adjusted for age and sex only using the same eight BMKs (CAMK1D, IFNAR1, GAB2, MDM2, KCNJ15, PRKCB, LYN, and PDE8A), yielded an AUC ROC curve of 0.911 (CI95%:[0.854-0.968]) on the test dataset with 80.0% sensitivity and 82.4% specificity (Figure 22B). Both models allow for clear separation of patients with depression from the control group.

[0113] [Table 9]

[0114] Analyzing the editing sites of the entire population and combining 17 editing variants of the five BMKs yielded an AUC of 0.751 (Sp 70.6%, Se 72.4%, Figure 6A). Using the eight biomarkers included in the multiplex and combining 53 differentially edited sites, the AUC increased to 0.824 with 76.2% specificity and 81% sensitivity (Figure 6B). Further increasing the number of editing sites of these eight BMKs to reach 79 sites, and sequencing all in the multiplex in a single experiment, the AUC increased further to 0.899 (Sp 82.5%, Se 81.7%, Figure 6C). The equation used in the latter case is as follows: Z=-(-0.181885362537332)*MDM2_Site68821586_W-(0.45875321689627)*MDM2_SWX+(0.932293378426399)*PRKCB_Site23920910_J-(0.235744342326864)*MDM2_W-(0.188045930057369)*MDM2_Site68821643+(0.129971821791191)*PRKCB_DIJK-(0.112543222902224)*MDM2_SVW+(0.0714766020787205)*CAMK1D_E-(-0.0566149561843161)*MDM2_SW+(0.343458398647021)*PRKCB_B-(0.0511754835571903)*GAB2_Site78331054+(0.0777964634309061)*PRKCB_Site23920911+(-0.309760257245042)*PRKCB_Site23920919_K+(0.0892196021964463)*MDM2_T+(0.0591702551145443)*PRKCB_J-(0.163034043783367)*KCNJ15_CE-(-0.3554615992579)*MDM2_Site68821616_X+(0.194383441652219)*CAMK1D_AE-(-0.0357242710579534)*KCNJ15_C+(-0.137804039545265)*PRKCB_JK-(0.073268779527063)*GAB2_CE-(0.583431075372027)*PDE8A_Site85096687_B-(0.083656284732413)*KCNJ15_Site38287529_C+(0.241597825714507)*PRKCB_Site23920902_I+(0.0897642626007787)*PRKCB_DIJ+(0.311212124874424)*PRKCB_Site23920893_G-(-0.0694365074403336)*KCNJ15_Site38287586-(0.356961148091174)*KCNJ15_Site38287525_B-(0.0284155839590126)*MDM2_Site68821573+(0.161274457136425)*IFNAR1.1_Site33355809-(-0.578010271214577)*PDE8A_B+(0.00936832259402957)*PRKCB_Site23920921_L-(-0.0220552696975651)*MDM2_WX+(0.0416981587266934)*IFNAR1.1_Site33355817-(0.056383762969179)*KCNJ15_Site38287589_E+(0.034447202518999)*IFNAR1.1_Site33355838_P-(0.181855602464856)*KCNJ15_Site38287590+(-0.232428154178395)*PRKCB_Site23920884_D-(0.0566507858081719)*KCNJ15_Site38287530_D+(0.00350968357773277)*PRKCB_IJK-(0.0986696717892891)*KCNJ15_Site38287591+(0.0416357072635886)*PRKCB_Site23920889_F+(-0.510062501581064)*PRKCB_Site23920844_B+(-0.201713338082801)*PRKCB_DJ-(0.219330533848547)*LYN_AC+(-0.0902753379195889)*IFNAR1.1_Site33355798+(0.0364611306918377)*PRKCB_DHIJK+(-0.0154244319603596)*GAB2_BCDE+(-0.0721590982562632)*IFNAR1.1_Site33355801_N+(0.0351216920222297)*CAMK1D_A+(0.130383083484889)*LYN_AG+(0.136853965680293)*GAB2_Site78331030-(0.201599671305856)*CAMK1D_Site12378440+(-0.0494625908197476)*IFNAR1.1_Site33355793_G-(0.112438468962481)*MDM2_KSX+(0.129569099073175)*IFNAR1.1_Site33355839_L-(0.12561875043809)*GAB2_ABDE-(0.127875837 752046)*PRKCB_DJM-(-0.421356734328941)*KCNJ15_ACE-(0.125572361520418)*MDM2_ASV-(-0.0421188089161209)*MDM2_Site68 821576_V+(0.333082982874752)*LYN_A+(0.205974310972437)*IFNAR1.1_Site33355762_F-(0.26063881758559)*MDM2_Site6882 1602-(0.16091158013623)*LYN_Site56012491-(0.585856554487505)*PRKCB_Site23920927_M-(-0.328951375049536)*KCNJ15_Si te38287526_A+(0.0681422708577013)*GAB2_EJ+(0.0265646151191535)*PRKCB_I+(-0.114728592530574)*LYN_Site56012490+(0 .0423564487829415)*CAMK1D_Site12378387+(-0.270495232628549)*IFNAR1.1_Site33355807_B-(0.202529212838335)*PRKCB_Si te23920952-(-0.0881700257476326)*KCNJ15_ACD+(0.0213312643325165)*GAB2_DE+(-0.00540846808767115)*MDM2_STX+(0.185 10874500411)*GAB2_Site78331008_D-(0.060213548631588)*MDM2_Site68821567-(-0.123438884158084)*MDM2_Site68821612_A.

[0115] Table 6 shows a list of combinations tested across the entire validation cohort (n=411). All combinations have a p-value < 10. -4The results were statistically significant between the control group and the cases.

[0116] [Table 10-1]

[0117] [Table 10-2]

[0118] [Table 10-3]

[0119] [Table 10-4]

[0120] [Table 10-5]

[0121] [Table 10-6]

[0122] [Table 10-7]

[0123] [Table 10-8]

[0124] Therefore, the biomarker panel analyzed above demonstrated excellent diagnostic performance for diagnosing DEP across the entire validation cohort (411 subjects), suggesting that the biomarker panel may indeed have clinical applicability for diagnosing DEP.

[0125] Next, we attempted to investigate whether the diagnostic performance of the assay could be further improved by considering sex-specific differences. In fact, the existence of sex-specific transcriptional signatures has been suggested for DEP. Previous studies have shown that the brain transcriptional profiles of DEP differ significantly between sexes, suggesting that the underlying molecular mechanisms of DEP may differ between women and men. 42 Therefore, RNA editing modifications were analyzed separately in the male and female populations of the validation cohort. Introducing sex specificity into the analysis resulted in a significant improvement in diagnostic performance. Notably, the female validation cohort (n=277) included more subjects than the male validation cohort (n=134), which reflects the proportion of men and women with DEP in the general global population. Indeed, in the female population of the validation cohort (n=277, 90 healthy controls and 187 DEP patients), the AUC for combining 11 RNA editing variants representing four different BMKS was 0.784 (Sp 71.1%, Se 78.6%, Figure 7A). The equation used in this case is as follows: Z=-(-0.000533736120198011)*MDM2_Site68821586_W-(0.211359639047036)*MDM2_W-(0.14626445233 2436)*MDM2_SWX-(0.160086235452429)*GAB2_Site78331054-(0.090792634422732)*MDM2_SW+(0.11767 2779615208)*MDM2_T+(0.209398709260439)*CAMK1D_AE+(0.115253600910406)*PRKCB_J-(0.193553830 436994)*CAMK1D_BC+(0.0404496406797322)*CAMK1D_E+(0.0771184775348071)*PRKCB_Site23920910_J

[0126] Increasing the number of biomarkers to six different targets and including 30 RNA editing variants increased the AUC to 0.835 (Sp 80.0%, Se 83.4%, Figure 7B). Furthermore, using the eight biomarkers included in the multiplex and combining 49 differentially edited sites, the AUC reached 0.880 (Sp 83.3%, Se 88.2%, Figure 7C), with both exhibiting over 80% specificity and sensitivity.

[0127] Machine learning approaches were used in both mROC and random forest (see Materials and Methods) to randomly divide sex-specific populations into training and test sets. When the female population was randomly divided into a training set (75% of the total population) and a test set (25% of the total population) and the above methods were applied, the test set using the mROC method yielded an AUC of 0.843 (Sp 88.9%, Se 80.4%, Figure 7D), and the MultDS random forest method yielded an AUC of 0.895 (Sp 86%, Se 80%, Figure 7E). Overall, the data obtained by the two different machine learning methods in the female validation cohort confirmed the superior diagnostic performance of the biomarker panel.

[0128] Table 16 shows a list of combinations tested in the female validation cohort (n=277). All combinations had a p-value < 10. -4 The results were statistically significant between the control group and the cases.

[0129] [Table 11-1]

[0130] [Table 11-2]

[0131] [Table 11-3]

[0132] [Table 11-4]

[0133] [Table 11-5]

[0134] [Table 11-6]

[0135] [Table 11-7]

[0136] Performance in the male population was even better. In fact, when the male population of the validation cohort (n=134, 53 healthy controls and 81 DEP patients) was analyzed separately, including 8 edited variants for 6 BMKs, an AUC of 0.831 was obtained (Sp 83.0%, Se 75.3%, Figure 8A). The equation used in this case is as follows: Z=-(0.238400918804649)*GAB2_CH-(0.230156347297484)*CAMK1D_DS-(0.1915 4950723228)*KCNJ15_Site38287525_B-(0.21914392389275)*GAB2_CE-(0.15578 3908162178)*LYN_Site56012491-(0.268738279594405)*MDM2_SVW-(0.1070369 76834934)*MDM2_Site68821643+(0.400819265369961)*PRKCB_Site23920893_G.

[0137] For seven biomarkers containing 39 RNA editing variants, the AUC was greater than 0.9 (AUC 0.936, Spe 92.5%, Se 87.7%, Figure 8B). Combining eight biomarkers included in the panel with these 68 editing variants of BMKS yielded a superior AUC of 0.976 (Spe 98.1%, Se 93.8%, Figure 8C). Using machine learning, the male population was randomly split into a training set (70% of the total population) and an experiment set (30% of the total population). The AUC obtained by mROC was 0.859 (Sp 87.5%, Se 87.5%, Figure 8D), while the random forest method yielded an AUC of 0.875 (Sp 82%, Se 80%, Figure 8E).

[0138] Table 8 shows a list of combinations tested in the male validation cohort (n=134). All combinations had a p-value < 10. -4 The results were statistically significant between the control group and the cases.

[0139] [Table 12-1]

[0140] [Table 12-2]

[0141] [Table 12-3]

[0142] [Table 12-4]

[0143] [Table 12-5]

[0144] [Table 12-6]

[0145] Therefore, the data obtained from both methods demonstrated excellent diagnostic performance for sex-specific diagnosis of DEP, with an AUC of over 0.850, specificity and sensitivity of over 80% in both women and men, which are indicators of a diagnostic assay that can be used clinically.

[0146] Having identified RNA editing biomarkers for the diagnosis of DEP, we attempted to investigate whether unedited isoforms of the validated targets could be useful in distinguishing controls from the patient population and whether they could be used in combination with the RNA editing biomarkers. We then calculated the AUC ROC for each unedited target, along with the significance of each unedited isoform. Analysis of the unedited isoforms of PRKCB, MDM2, IFNAR1, and PDE8A in the validation cohort (n=411 subjects) yielded AUCs greater than 0.5, and both the unedited isoforms of PRKCB and MDM2 were significantly different between controls and DEP (p<0.05, Table 9).

[0147] [Table 13]

[0148] In addition, combining unedited isoforms with a single isoform of the same target revealed significant differences between healthy controls and DEP patients (Table 19), suggesting that unedited isoforms can also be used in diagnostic assays.

[0149] [Table 14]

[0150] conclusion To the best of our knowledge, this study is the first to investigate RNA editing modifications used as blood biomarkers for diagnosing mood disorders. We conducted an editome analysis of DEP in patients recruited from the outpatient department of the Department of Emergency Psychiatry and Post-Acute Care (CHRU of Montpellier). The presence of a psychiatric diagnosis was confirmed by trained psychiatrists administering the French version of the Montgomery-Asberg Depression Rating Scale (MADRS) and the Clinician Assessment (IDS-C30), a 30-item Inventory of Depressive Symptomatology, for scoring depression. We explored the association between depression and patient editomes according to RNA-Seq studies in a discovery cohort of 64 subjects. We identified 646 A-to-I edit variants representing 366 genes that were differentially edited between depressed patients and healthy controls. This gene set demonstrates potent enrichment of biological processes involving multiple immune categories, thus confirming potent immune components in the pathophysiology of DEP. 43 Next, the list of variants was narrowed down to a subset of 14 genes (17 variants) by using very stringent criteria, and these biomarkers were individually validated in a prioritized cohort of 86 individuals. These genes also showed strong enrichment in genes associated with the regulation of immune system processes. In addition, some were strongly associated with pathways involved in the nervous system and neurotransmitter receptor-related pathways. Finally, different combinations of the panel of eight markers were identified, and these combinations were validated in a large cohort of 411 patients. Furthermore, by introducing sex specificity and using machine learning approaches, the performance of the diagnostic assays was substantially increased. The results identified a combination of biomarkers with high diagnostic performance that is clinically useful and preferable for objectively diagnosing depressive disorders from blood samples.

[0151] Previously, capillary electrophoresis single-strand conformation 5 polymorphism (CE-SSCP) has identified alterations in serotonin receptor 5HT2CR and PDE8A mRNA editing in brain samples from the descendants of suicide victims. 5、44 In recent years, alterations in RNA editing have been demonstrated in psychotic patients with chronic hepatitis C virus (HCV) receiving IFN-α and ribavirin combination antiviral therapy, where treatment-induced depression is a common complication. 22 In later studies, the measurement of editing levels is based on targeted next-generation sequencing (NGS), which provides both high throughput and sufficient depth per base, enabling reliable quantification at all sites studied.

[0152] References 1.American Psychiatric Association.Diagnostic and Statistical Manual of Mental Disorders.5th ed.Washington,DC.2013. 2.Dunlop K, Talishinsky A, Liston C.Intrinsic Brain Network Biomarkers of Antidepressant Response:a Review.Current psychiatry reports.Aug 13 2019;21(9):87. 3.Jentsch MC, Van Buel EM, Bosker FJ, et al.Biomarker approaches in major depressive disorder evaluation in the context of current hypotheses.Biomarkers in medicine.2015;9(3):277-297. 4. Smith KM, Renshaw PF, Bilello J. The diagnosis of depression: current and emerging methods. Comprehensive psychiatry. Jan 2013;54(1):1-6. 5.Chimienti F,Cavarec L,Vincent L,et al.Brain region-specific alterations of RNA editing in PDE8A mRNA in suicide decedents.Translational psychiatry.Feb 15 2019;9(1):91. 6.Dobin A,Gingeras TR.Optimizing RNA-Seq Mapping with STAR.Methods in molecular biology.2016;1415:245-262. 7.DePristo MA,Banks E,Poplin R,et al.A framework for variation discovery and genotyping using next-generation DNA sequencing data.Nature genetics.May 2011;43(5):491-498. 8.John D,Weirick T,Dimmeler S,Uchida S.RNAEditor:easy detection of RNA editing events and the introduction of editing islands.Briefings in bioinformatics.Nov 1 2017;18(6):993-1001. 9.Yang H,Wang K.Genomic variant annotation and prioritization with ANNOVAR and wANNOVAR.Nature protocols.Oct 2015;10(10):1556-1566. 10.Tempel S.Using and understanding RepeatMasker.Methods in molecular biology.2012;859:29-51. 11.Kasprzyk A,Keefe D,Smedley D,et al.EnsMart:a generic system for fast and flexible access to biological data.Genome research.Jan 2004;14(1):160-169. 12.Langmead B,Salzberg SL.Fast gapped-read alignment with Bowtie 2.Nature methods.Mar 4 2012;9(4):357-359. 13.Li H,Handsaker B,Wysoker A,et al.The Sequence Alignment / Map format and SAMtools.Bioinformatics.Aug 15 2009;25(16):2078-2079. 14.Li H.A statistical framework for SNP calling,mutation discovery,association mapping and population genetical parameter estimation from sequencing data.Bioinformatics.Nov 1 2011;27(21):2987-2993. 15.Gentleman RC,Carey VJ,Bates DM,et al.Bioconductor:open software development for computational biology and bioinformatics.Genome biology.2004;5(10):R80. 16.Linden A.Measuring diagnostic and predictive accuracy in disease management:an introduction to receiver operating characteristic(ROC)analysis.Journal of evaluation in clinical practice.Apr 2006;12(2):132-139. 17.Kramar A,Faraggi D,Fortune A,Reiser B.mROC:a computer program for combining tumour markers in predicting disease states.Computer methods and programs in biomedicine.Sep 2001;66(2-3):199-207. 18.Breiman L.Random Forests.Machine Learning.Kluwer Academic Publishers.2001;45:5-32. 19.Consortium GO.Gene Ontology Consortium:going forward.Nucleic acids research.Jan 2015;43(Database issue):D1049-1056. 20.Raudvere U,Kolberg L,Kuzmin I,et al.g:Profiler:a web server for functional enrichment analysis and conversions of gene lists(2019 update).Nucleic acids research.Jul 2 2019;47(W1):W191-W198. 21.Pomaznoy M,Ha B,Peters B.GOnet:a tool for interactive Gene Ontology analysis.BMC bioinformatics.Dec 7 2018;19(1):470. 22.Salvetat N,Van der Laan S,Vire B,et al.RNA editing blood biomarkers for predicting mood alterations in HCV patients.Journal of neurovirology.Jul 22 2019. 23.Szklarczyk D,Gable AL,Lyon D,et al.STRING v11:protein-protein association networks with increased coverage,supporting functional discovery in genome-wide experimental datasets.Nucleic acids research.Jan 8 2019;47(D1):D607-D613. 24.Shelton RC.The molecular neurobiology of depression.The Psychiatric clinics of North America.Mar 2007;30(1):1-11. 25.Costas J,Gratacos M,Escaramis G,et al.Association study of 44 candidate genes with depressive and anxiety symptoms in post-partum women.Journal of psychiatric research.Aug 2010;44(11):717-724. 26.Colledge M,Snyder EM,Crozier RA,et al.Ubiquitination regulates PSD-95 degradation and AMPA receptor surface expression.Neuron.Oct 30 2003;40(3):595-607. 27.Liu B,Mink S,Wong KA,et al.PIAS1 selectively inhibits interferon-inducible genes and is important in innate immunity.Nature immunology.Sep 2004;5(9):891-898. 28.Murakami Y,Ishibashi T,Tomita E,et al.Depressive symptoms as a side effect of Interferon-alpha therapy induced by induction of indoleamine 2,3-dioxygenase 1.Scientific reports.Jul 20 2016;6:29920. 29.Brodie EJ,Infantino S,Low MSY,Tarlinton DM.Lyn,Lupus,and(B)Lymphocytes,a Lesson on the Critical Balance of Kinase Signaling in Immunity.Frontiers in immunology.2018;9:401. 30.Hayashi T,Umemori H,Mishina M,Yamamoto T.The AMPA receptor interacts with and signals through the protein tyrosine kinase Lyn.Nature.Jan 7 1999;397(6714):72-76. 31.Moura-Alves P,Fae K,Houthuys E,et al.AhR sensing of bacterial pigments regulates antibacterial defence.Nature.Aug 28 2014;512(7515):387-392. 32.Song MS,Song SJ,Kim SY,Oh HJ,Lim DS.The tumour suppressor RASSF1A promotes MDM2 self-ubiquitination by disrupting the MDM2-DAXX-HAUSP complex.The EMBO journal.Jul 9 2008;27(13):1863-1874. 33.Wang H,Nestor CE,Benson M,Zhang H.GAB2 regulates type 2 T helper cell differentiation in humans.Cytokine.Aug 2017;96:234-237. 34.Verploegen S,Lammers JW,Koenderman L,Coffer PJ.Identification and characterization of CKLiK,a novel granulocyte Ca(++) / calmodulin-dependent kinase.Blood.Nov 1 2000;96(9):3215-3223. 35.Lin E,Kuo PH,Liu YL,Yu YW,Yang AC,Tsai SJ.A Deep Learning Approach for Predicting Antidepressant Response in Major Depression Using Clinical and Genetic Biomarkers.Frontiers in psychiatry.2018;9:290. 36.Zhou X,Chen Y,Mok KY,et al.Identification of genetic risk factors in the Chinese population implicates a role of immune system in Alzheimer’s disease pathogenesis.Proceedings of the National Academy of Sciences of the United States of America.Feb 20 2018;115(8):1697-1706. 37.Beurel E,Lowell JA.Th17 cells in depression.Brain,behavior,and immunity.Mar 2018;69:28-34. 38.Fischer EH,Charbonneau H,Tonks NK.Protein tyrosine phosphatases:a diverse family of intracellular and transmembrane enzymes.Science.Jul 26 1991;253(5018):401-406. 39.Irie-Sasaki J,Sasaki T,Matsumoto W,et al.CD45 is a JAK phosphatase and negatively regulates cytokine receptor signalling.Nature.Jan 18 2001;409(6818):349-354. 40.Basterzi AD,Yazici K,Buturak V,et al.Effects of venlafaxine and fluoxetine on lymphocyte subsets in patients with major depressive disorder:a flow cytometric analysis.Progress in neuro-psychopharmacology & biological psychiatry.Feb 1 2010;34(1):70-75. 41.Aston C,Jiang L,Sokolov BP.Transcriptional profiling reveals evidence for signaling and oligodendroglial abnormalities in the temporal cortex from patients with major depressive disorder.Molecular psychiatry.Mar 2005;10(3):309-322. 42.Seney ML,Huo Z,Cahill K,et al.Opposite Molecular Signatures of Depression in Men and Women.Biological psychiatry.Jul 1 2018;84(1):18-27. 43.Wohleb ES,Franklin T,Iwata M,Duman RS.Integrating neuroimmune systems in the neurobiology of depression.Nature reviews.Neuroscience.Aug 2016;17(8):497-511. 44.Weissmann D,van der Laan S,Underwood MD,et al.Region-specific alterations of A-to-I RNA editing of serotonin 2c receptor in the cortex of suicides with major depression.Translational psychiatry.Aug 30 2016;6(8):e878.

Claims

1. An in vitro method for assisting the diagnosis of depressive disorder in human patients from biological samples of the said patients, wherein the method is a) For combinations of at least two A to I edited RNA biomarkers, - For each selected biomarker, the relative proportion of RNA editing at a given editing site to at least one or a combination of sites that can be edited in the RNA transcripts of the at least two biomarkers, and / or - Determining the relative percentage of one isoform or combination of isoforms of each of the two biomarkers' RNA transcripts, The determination that at least two A to I edited RNA biomarkers are selected from the group consisting of PRKCB, MDM2, PIAS1, IFNAR1, IFNAR2, LYN, AHR, RASSF1, GAB2, CAMK1D, KCNJ15, IL17RA, PTPRC, and PDE8A biomarkers, b) Determining a final result value based on an algorithm or equation that includes the result values ​​obtained for each of the at least two biomarkers and the result values ​​obtained for each selected biomarker, c) Determining the final result value or depression score obtained for the patient, and the control result value obtained for a healthy control subject, wherein the control result value is determined in a manner equivalent to that of the final result value. d) A method comprising: classifying the patient as positive or having a depressive disorder if the final result value or depression score value is higher than a threshold / cutoff; or classifying the patient as negative or not having a depressive disorder if the final result value is not higher than the threshold.

2. The method according to claim 1, wherein in step b), the at least two AtoI edited RNA biomarkers are selected from the group consisting of PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A.

3. In step b), the result value or depression score is calculated by an algorithm that implements a multivariate method, and the multivariate method is - An mROC program for maximizing AUC (Area Under Curve) ROC, providing equations for each combination and identifying linear combinations that can be used as the following new virtual marker Z, Z = a. (Biomarker 1) + b. (Biomarker 2) + ... i. (Biomarker i) + ... n. (Biomarker n) In the formula, i is the calculated coefficient and (biomarker i) is the level of the biomarker to be considered, mROC program and / or - A random forest (RF) approach and / or selectively applied to evaluate the combination of RNA editing sites and / or isoforms, to rank the importance of the RNA editing sites and / or isoforms, and to combine the best RNA editing sites and / or isoforms. - A multivariate analysis applied to evaluate the combination of RNA editing sites and / or isoforms for assisting the diagnosis, wherein the multivariate analysis is, for example, - Logistic regression models, and / or, that can be applied to univariate and multivariate analyses to estimate the relative risk of a patient at different levels of RNA editing sites or isoform values. - The CART (Classification and Regression Tree) approach applied to evaluate combinations of RNA editing sites and / or isoforms, and / or - Support Vector Machine (SVM) approach, - Artificial Neural Network (ANN) approach, - Bayesian network approach, - WKNN (Weighted k-Nearest Neighbors) approach, - Partial Least Squares Discriminant Analysis (PLS-DA), - Linear and quadratic discriminant analysis (LDA / QDA), and The method according to claim 1 or 2, comprising multivariate analysis selected from the group consisting of any other mathematical methods combined with biomarkers.

4. The method according to any one of claims 1 to 3, wherein the depressive disorder is DEP.

5. The method according to any one of claims 1 to 4, wherein the biological sample is whole blood, serum, plasma, urine, cerebrospinal fluid, or saliva.

6. The method according to claim 5, wherein the biological sample is a whole blood sample.

7. The method according to any one of claims 1 to 6, wherein for each selected biomarker, the result value is statistically / specifically different from the control result value by p < 0.

05.

8. Regarding the biomarker or combination of biomarkers selected in step a), - Coverage > 30, - AUC > 0.8, -0.95 > Magnification change > 1.05, and The method according to any one of claims 1 to 7, wherein the criterion -p < 0.05 must be satisfied.

9. The selected A to I edited RNA biomarker included in step a) The method according to any one of claims 1 to 8, selected from the group consisting of combinations of at least 3, 4, 5, 6, 7, or 8 of the following biomarkers: PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A.

10. To evaluate the combination of RNA editing sites and / or isoforms, to rank the importance of the RNA editing sites and / or isoforms, and to combine the best RNA editing sites and / or isoforms, a random forest (RF) algorithm is applied, and if the model is adjusted for age, sex, psychiatric treatment, and addiction, or if the model is adjusted for age and sex, the selected A to I edited RNA biomarkers include, in step a), a combination of the following eight biomarkers: GAB2, IFNAR1, KCNJ15, LYN, MDM2, PRKCB, CAMK1D, and PDE8A, and the calculation of the relative percentage of at least one of the RNA editing sites or isoforms is shown in Tables 1 and 2 below. Table 1 Table 2 The method according to any one of claims 1 to 9, based on the RNA editing site or isoform listed therein.

11. The above combination of at least two, three, four, five, six, seven, or eight selected biomarkers is, for each of the selected biomarkers, Table 2A (edited area), Table 3 Table 2B (Isoforms), Table 4 Or Table 5 (edited section) Table 5 The method according to any one of claims 1 to 10, comprising the calculation of the relative percentage of at least one of the RNA editing sites or isoforms listed above.

12. The method according to any one of claims 2 to 11, wherein the combination of at least two, three, four, five, six, seven, or eight selected biomarkers is selected from the combinations comprising the following eight biomarkers: PRKCB, MDM2, IFNAR1, KCNJ15, LYN, CAMK1D, GAB2, and PDE8A.

13. An in vitro method for assisting in the diagnosis of depressive disorder in a human patient according to any one of claims 1 to 12, wherein, if the patient being tested is female, the biomarker of the combination of at least two single A to I edited RNA biomarkers in step a) is selected from the group of PRKCB, MDM2, PIAS1, IFNAR1, IFNAR2, LYN, AHR, RASSF1, GAB2, CAMK1D, KCNJ15, IL17RA, PTPRC, and PDE8A biomarkers.

14. An in vitro method for assisting the diagnosis of depressive disorder in a human patient according to any one of claims 1 to 12, wherein, if the patient being tested is male, in step a) the biomarker of the combination of at least two single A to I edited RNA biomarkers is selected from the group of biomarkers consisting of the combination of PRKCB, MDM2, PIAS1, IFNAR1, IFNAR2, LYN, AHR, RASSF1, GAB2, CAMK1D, KCNJ15, IL17RA, PTPRC, and PDE8A biomarkers.

15. An in vitro method for assisting the diagnosis of depressive disorder in a human patient, wherein the method comprises a method for assisting the diagnosis of depression in a human patient according to any one of claims 1 to 14, wherein the combination of biomarkers associated with the combination used to assist in the diagnosis of depressive disorder and / or the Z equation differs depending on whether the patient being tested is female or male.

16. The method according to any one of claims 1 to 15, wherein in step a), the relative proportion of the RNA edit and / or the percentage of the isoform in a given edit is measured in the biological sample by NGS.

17. In step a), the amplicon / nucleic acid sequence used for the detection of the RNA editing site and / or isoform of the RNA transcript of the biomarker is - For each of the selected biomarkers, a set of primers (SEQ ID NOs. 1 to 36) listed in Table 1, or - The method according to any one of claims 1 to 16, wherein a set of primers or a combination of primers that enables obtaining an amplicon, the set of primers or a combination of primers that includes or is identical to the amplicon obtainable by the set of primers listed in Table 1 (SEQ ID NOs: 1 to 36).

18. A kit for determining whether a patient exhibits depressive disorder, wherein the kit is 1) Instructions for obtaining the result value or depression score by optionally applying the method of any one of claims 1 to 17, wherein the analysis of the result value or depression score determines whether the patient exhibits a depressive disorder. 2) i) A combination of at least two sets of primers from a group of primer pairs selected from Sequence ID No. 1 to Sequence ID No. 36, wherein the selection depends on the combination of biomarkers used to aid in the diagnosis of depressive disorders, or ii) a combination of at least two sets of primers that enables the acquisition of an amplicon, the combination of at least two sets of primers being identical to the amplicon obtained by the at least two sets of primers selected in i).

Images