Pharmacogenic decision-making support for NMDA, glycine, and AMPA receptor modulators

Abstract

To provide a method and a computing device that allow for more effective determination regarding which patients will experience drug efficacy, and which patients will experience adverse drug events, and that provide personalized recommendations for patients regarding dose, the frequency of drug administration, and drug choice.SOLUTION: A method for identifying patients diagnosed with treatment resistant or refractory depression, pain or other clinical indications, who are eligible to receive N-methyl-D-aspartate receptor antagonist, glycine receptor beta modulator, or α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor-based therapies.SELECTED DRAWING: None

Description

Technical Field

[0001] Cross - reference to related applications This application claims priority to and the benefit of the filing date of (1) U.S. Provisional Application No. 62 / 795,705, filed on January 23, 2019, entitled "Methods and Systems for Reconstructing a Drug Space Network from Pharmacogenomic Regulatory Interactions, and Their Use", and (2) U.S. Provisional Application No. 62 / 795,710, filed on January 23, 2019, entitled "Companion Diagnostic Assays for N - Methyl - D - Aspartate Receptor Modulators", the entire disclosure of each of which is hereby expressly incorporated by reference.

[0002] The technology described herein relates to pharmacogenomic clinical decision - support assays useful for the selection of therapies for N - methyl - D - aspartate (NMDA) receptor, glycine receptor, and α - amino - 3 - hydroxy - 5 - methyl - 4 - isoxazolepropionic acid (AMPA) receptor - based depression (particularly treatment - resistant or refractory depression), as well as other clinical indications including anesthesia and analgesia / anesthesia - pain disorders, neuropsychiatric disorders, and neurological disorders, using ketamine and its enantiomers as examples. More specifically, these technologies relate to specific biomarkers including gene markers, clinical value, and disease phenotypes derived from patients, for optimizing the selection of drugs that affect these receptor networks and dosages in individual patients.

Background Art

[0003] Existing antidepressants are ineffective for many patients. A new class of antidepressants targeting glutamate receptors in the human forebrain is being developed. Ketamine (RS-2-chlorophenyl-2-methylaminocyclohexanone), a glutamate N-methyl-d-aspartate receptor (NMDAR) non-competitive antagonist approved by the U.S. Food and Drug Administration (FDA) as an anesthetic, has shown promise as an antidepressant for patients with treatment-resistant depression (TRD). While the racemic formula can have potent and undesirable psychotropic and other side effects depending on several variables, ketamine's chemical analogues show reduced adverse events. Intravenous and oral formulations have demonstrated efficacy and tolerability in controlled and open-label trials across a patient population that is well known to often have little to no response to conventional antidepressants targeting the serotonin transporter (also known as SLC6A4, 5HTT, or SERT1), including serotonin-norepinephrine reuptake inhibitors (SNRIs). The evidence suggests that ketamine, its enantiomers, and ketamine analogs exert their mechanisms of action primarily through the modulation of NMDA receptors (NMDARs) and downstream receptors within this network in the human brain.

[0004] The pharmacodynamic (PD) targets of ketamine-like drugs are NMDARs consisting of GRIN1 and GRIN2 subunits, which bind to glutamate and N-methyl-D-aspartate, glycine and D-serine binding sites encoded by GLRBs, and sites that bind to polyamines, histamine and cations. Antagonists, partial antagonists, and receptor modulators (such as ketamine and its NMDARs and other NMDARs), as well as glycine modulators (including phencyclidine, amantadine, dextromethorphan, tyrethamine, riluzole, methoxetamine, metoxphenidine, and memantine), bind to inward-directed Ca +2Blocking inflow prevents postsynaptic depolarization. Neuroimaging studies have demonstrated that intravenous infusion of ketamine induces a transient surge in glutamate levels observed in the prefrontal cortex, in conjunction with a rapid antidepressant effect. Following NMDAR blockade, glutamate has been shown to preferentially bind to the GRIA1, GRIA2, and GRIA4 subunits of the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor. Several NMDAR antagonists and partial antagonists, GLRB modulators, and AMPAR agonists are under development for the treatment of treatment-refractory depression, but exhibit dissociative effects in patients. Furthermore, NMDAR antagonists, GLRB antagonists, and AMPAR modulators, partial antagonists, and receptor modulators that act in the same network as this class of drugs that have failed in clinical trials due to safety concerns may be more likely to succeed in clinical trials by being repurposed using the methods of this disclosure.

[0005] The three-dimensional (3D) architecture of the human regulatory epigenome plays a major role in determining the human phenotype. It is now recognized that the majority of key single nucleotide polymorphisms (SNPs) associated with disease risk, drug response, and other human traits, as found in genome-wide association studies (GWAS), reside within enhancers, promoters, and other non-coding regulatory elements. Coupled with recent insights into the construction of functional 4D nucleomes, the abundance of biomedical "big data" accessible in open-source and proprietary resources has enabled the reconstruction of drug regulatory pathways in the human genome. Novel methods are being applied to mining regulatory variants from the results of genome-wide association studies (GWAS), phenome-wide association studies (PheWAS), and mining of electronic health record and clinical trial data. It is now common knowledge in the art that the association of SNP characteristics from these data sources is crucial because it can be re-evaluated from a pathway analysis perspective and resolved into the same or related biological networks. By combining extensive biomedical data with innovative methods for mining biological networks, a foundation is provided for detecting gene variants that may influence the variability of drug responses, including adverse drug events. This approach promises significant advances in specialties such as psychiatry, where a lack of drug efficacy and a high incidence of adverse drug events have proven to be particularly problematic in patient care.

[0006] More than half of all Americans experience symptoms of a mental disorder at some point in their lives. The most common lifelong mental disorders are stress and anxiety disorders, mood disorders (including major depressive disorder and bipolar disorder), impulse control disorders, substance use disorders, and schizoaffective disorder. The lifetime prevalence of any one mental disorder in the United States is 53%, with 28% having two or more lifetime disorders and 18% having three or more lifetime disorders, indicating that comorbid mental disorders are a significant medical challenge. Some mental disorders, such as bipolar I disorder, are inherited within families with a penetrant rate of approximately 80%, while other mental disorders do not show clear heritability. In the case of major depressive disorder (MDD), genetic factors play a significant role in the pathogenesis of the disease, as shown by family, twin, and adoption studies. Twin studies suggest a 50% heritability, and family studies show that the lifetime risk of developing MDD is two to three times higher among first-degree relatives. Several sociodemographic variables are significantly associated with lifetime risk of mental disorders in cohort-controlled studies. For example, women (biological sex) have a significantly higher risk of anxiety disorders and major depressive disorder than men, and men (biological sex) have a significantly higher risk of impulse control disorders and substance abuse disorders than women. Non-Hispanic Black and Hispanic White individuals have a significantly lower risk of anxiety, mood, and substance abuse disorders than non-Hispanic White individuals, and lower education levels are associated with a higher risk of substance abuse disorders. The data show that many factors, ranging from environmental and sociological factors to biological sex, ethnicity, and familial genetics, all contribute to the etiology of mental disorders. Furthermore, the complexity of phenotypes in individual patients or within patient cohorts, including comorbidities with other mental disorders and stress-related disorders, necessitates a range of different algorithmic classification solutions, including machine learning, and multiple statistical analyses, including linear regression, to accurately identify precision therapies beyond those currently available.

[0007] Mental illness has a greater impact on human health than any other disease. For example, major depressive disorder (MDD) causes a greater disability burden worldwide than any other medical condition, such as cancer, heart disease, stroke, chronic obstructive pulmonary disease, and HIV / AIDS, yet it remains the most undiagnosed, misdiagnosed, untreated, or poorly treated disease known to humankind. In 2013, the U.S. National Institutes of Health (NIH) allocated 13 times more funding to oncology research than to depression research (approximately $5.3 billion compared to $415 million). In the United States, from 2009 to 2011, adverse events associated with prescribed antidepressants reached more than 25,000 emergency room visits annually, accounting for 30% of all prescription drug-related hospitalizations each year. Patients with major depressive disorder and comorbid medical conditions experience more severe depressive symptoms, as well as lower response and remission rates to antidepressant treatment, compared to patients without comorbid conditions. Treatment-resistant depression (TRD) accounts for 30–40% of all patients diagnosed with MDD and is defined as "failure to achieve remission after two established antidepressant courses known to be at evidence-based, acceptable doses and durations."

[0008] Modern antidepressants are ineffective for many patients, and even in those who respond or achieve remission, several weeks to months of medication are required before symptom relief is achieved. Therefore, newer and more effective antidepressants are being developed. For example, both racemic mixtures of ketamine and the S-enantiomer of ketamine are examples of N-methyl-D-aspartate receptor (NMDAR) partial antagonists approved by the U.S. Food and Drug Administration (FDA) for the treatment of TRD. Ketamine induces a rapid antidepressant response and associated elevation of cortical glutamate levels in approximately 50% of TRD patients, as measured by the total score on the Montgomery-Åsberg Depression Rating Scale (MADRS). R,S-ketamine has been used since 1970 for clinical indications such as chronic pain, perioperative analgesia, and sedation, but adverse drug events (AEs) after ketamine treatment are common, and its use is limited to inpatient and outpatient settings to restrict its repurposing. For example, in a Phase III clinical trial of esketamine for TRD prior to submission to the FDA, nearly a quarter of TRD patients experienced severe delusional symptoms, and two deaths were reported. A further 6.9% of TRD patients in the treatment group experienced severe psychotic effects such as delirium, delusions, suicidal ideation, and suicide attempts.

[0009] One of the challenges in psychiatry is the precise matching of pharmacotherapy to accurately address the complex symptoms of individual patients. Psychotic patients exhibit a wide range of comorbidities, and there are few objective biomarkers that can be used as diagnostic criteria to accurately tailor antidepressants, antipsychotics, and antimanic therapies to patients. While diagnostic rating scales such as the Hamilton Depression Scale (HAM-D) have shown high inter-rater reliability, mental disorders such as depression exist in a variety of distinct phenotypes. Non-pharmacological therapies may show improved efficacy in patients with TRD or relapsing depression. For example, repeated transcranial magnetic stimulation (rTMS) is a promising alternative to non-pharmacological therapies in patients with TRD. However, the best results in TRD are obtained when rTMS is used as an adjunct to conventional antidepressant pharmacotherapy, as is the case with antidepressant classes including NMDAR antagonists or partial antagonists, GLRB modulators, or AMPAR agonists, which can only be offered in clinical practice to patients already taking another antidepressant.

[0010] rTMS therapy requires dozens of clinical visits, remission rates vary greatly among TRD patients, and rTMS is only effective in about 20-40% of cases where remission from depression lasts for 1-2 years. Recent results from rTMS combined with neuroimaging demonstrate specific clustering of depressed patients into four distinct phenotypes along the anhedonia and anxiety neurosis axes, based on differences in rTMS array configuration. These results provide substantial evidence that stratifying psychotic patients by phenotype based on the activation of different brain connectivity pathways and networks exhibiting considerable inter-individual variability among patients can significantly improve the opportunity for precise matching of optimal therapies for individual patients.

[0011] While rTMS shows promise in TRD patients, its mechanism of action remained unclear until independent research demonstrated that TMS initially acts in the anterior cingulate cortex, significantly increasing glutamate levels along with biomarkers of N-methyl-d-aspartate receptor (NMDAR) modulation.

[0012] These findings are noteworthy because they demonstrate that the mechanism of brain activation by rTMS is substantially indistinguishable from the mechanism of action of ketamine pharmacotherapy. Therefore, while rTMS and ketamine exhibit similar mechanisms of action for alleviating TRD, approximately half of all TRD patients do not achieve remission after treatment with either treatment option. Furthermore, both rTMS and ketamine cause transient but serious adverse events (AEs), including dissociation (the presumptive basis for ketamine's analgesic effect), psychotic manifestations, and neurocognitive impairment. This suggests that it is crucial to match individual patients to one of these therapies or other antidepressants, if it is possible to select which patients will benefit from these treatments and which will unnecessarily suffer serious AEs without adequate antidepressant effect.

[0013] A recent study combining transcranial magnetic stimulation (TMS) for alleviating depression with subsequent neuroimaging demonstrated that TRD patients can be clearly stratified into four subtypes based on their response to TMS device placement, due to four distinct endogenous neuroanatomical pathways activated in association with distinctly different symptom clusters. These four subtypes were determined independently, as demonstrated in this disclosure, using a combination of clinical and molecular data, thereby providing an example for other psychiatric and stress-related disorders where improved pharmacophenomics decision support can offer better treatment options for patients. Similarly, NMDAR antagonist therapy can be used as an adjunct to age-related degenerative conditions. [Overview of the Initiative]

[0014] Various methods can be used to accurately determine the precise treatment requirements of individual patient phenotypes or phenotypic cohorts. This disclosure describes methods for constructing pharmacophenomics assays for clinical decision support or companion diagnostics of psychotropic drugs to optimize the fit of therapeutic interventions to individual patients or patient cohorts diagnosed with psychiatric or related disorders (such as treatment-resistant depression, chronic pain, migraine, fibromyalgia, inflammatory diseases, and other conditions in which ketamine or one analogue constitutes an effective treatment). In the context of this disclosure, a patient's drug response and adverse event phenotype consists of a set of multiple variables described herein, ranging from the patient's unique configuration of the drug's pharmacogenomics network, including its variant profile configuration, to behavioral phenotypes that may be derived from clinical data.

[0015] The methods used in this disclosure for patient stratification utilize different data sources, some of which may be incomplete, require data cleansing and / or curation, or may be nonexistent. The different methods described herein range from those that can accommodate different combinations of limited data to those that can accommodate broader computational solutions, or those that can bridge missing data elements using probabilistic methods.

[0016] This disclosure includes a set of interconnected, distinct methods for providing accurate pharmacofenomics decision support to patients diagnosed with mental disorders. The output provides quantitative scores for ranking therapeutic interventions, including recommendations for drug selection and dosage, transcranial magnetic stimulation, electroconvulsive therapy, and behavioral interventions. This disclosure includes pharmacofenomics methods for classifying patients diagnosed with mental disorders into subtypes for optimization of therapeutic interventions. These methods can be used to construct diagnoses that recommend the best therapeutic agreement for individual patients. In another embodiment, these methods can be used to enhance patient selection based on pharmacofenomics stratification prior to clinical trials. In another embodiment, these methods can be used to construct companion diagnoses for psychotropic drugs to ensure patient safety during drug development, marketing, and post-marketing.

[0017] In another embodiment, clinical values ​​are obtained from an EHR or similar source, and SNPs in the PD and PK genes are obtained from the patient's genotype; these are input as quantitative values ​​into a regression equation (nomogram) to determine the drug dosage of a drug such as ketamine for that individual patient. In this embodiment, the optimal therapeutic dose is developed in a stepwise regression model equation that includes genetic and clinical values, and the regression model is retested and validated using a patient population to ensure the accuracy of the output of the regression equation, which may be determined by a receiver operating curve, as the area under the curve (AUC).

[0018] In another embodiment, clinical values ​​are obtained from EHRs or similar sources, and SNPs in PD and PK genes are combined with disease risk SNPs obtained from genome-wide association studies (GWAS) and selectively annotated into subsystems specific to adverse events and efficacy in drug pharmacological genomics networks such as ketamine, thereby predicting whether a patient will benefit from a drug and, if so, determining an appropriate dosage for the patient.

[0019] In another embodiment, clinical values ​​are obtained from EHRs or similar sources, and SNPs of PD and PK genes are combined with disease risk SNPs obtained from GWAS and PheWAS, and selectively annotated into subsystems specific to adverse events and efficacy in drug pharmacological genomics networks such as ketamine, to predict whether a patient will benefit from a drug and, if so, determine an appropriate dosage for the patient. In this embodiment, therapeutic drug monitoring by pharmacological metabolomics is used to collect more accurate data on existing prescription and non-prescription drugs used by the patient and their metabolites through analysis of biological samples (blood, cheek swabs, urine or other body fluids) obtained from the patient or a cohort of patients.

[0020] In yet another embodiment, clinical values ​​are obtained from EHRs or similar sources, and SNPs of PD and PK genes are combined with disease risk SNPs obtained from GWAS and PheWAS to selectively annotate into pharmacogenetic adverse events and efficacy-specific subnetworks for drugs such as ketamine, and these data are matched to one of four phenotypes determined using Hamilton Depression Rating Scale (HAMD) scoring in terms of antidepressants such as ketamine.

[0021] In another embodiment, pharmacophenomics decision support is determined using inputs from a model that includes: (1) molecular profiling of drug-induced subnetworks in patients or cohorts of patients; (2) clinical variables obtained from electronic health records or equivalent measurements performed by clinicians; (3) patient subtyping based on clinical variables and neuroimaging studies; and (4) SNPs of PD and PK stratifying patients by drug response. Furthermore, drug-drug and drug-gene interactions objectively measured using drug metabolism methods can be used to minimize adverse drug events in individual patients or cohorts of patients.

[0022] Another embodiment of this system is a companion diagnostic configuration that can be used for patient selection for clinical trials between the commercial and post - commercial stages of drugs such as antidepressants that act as NMDAR antagonists, partial antagonists, GLRB modulators, and AMPAR agonists.

[0023] Another embodiment of this system is to determine and select the addition of a therapeutic agent to an NMDAR modulator in order to improve outcomes. Another embodiment of the methods and systems described herein can be used for re - evaluation of drugs and diversion of drugs for clinical trials.

Brief Description of the Drawings

[0024] [Figure 1A] A block diagram of a computer network and system in which an exemplary companion diagnostic system can operate according to the embodiments described herein is shown. [Figure 1B] A block diagram of an exemplary drug and dosage determination server that can operate in the system of FIG. 1A according to the embodiments described herein. [Figure 1C] A block diagram of an exemplary client device that can operate in the system of FIG. 1A according to the embodiments described herein. [Figure 1D] A flow diagram representing an exemplary method for determining the drug and dosage to be administered to a patient suffering from depression or other neuropsychiatric disorders based on comparing data from the patient's biological sample to a pharmacogenomic network specific to the reference drug and a constituent subnet of the drug of interest. [Figure 2] Examples of measurements collected from a patient's biological sample using chromosome conformation capture, bioinformatics analysis, and / or similar measurements are shown. [Figure 3]A simple example is shown of how an SNP within an enhancer in a network can disrupt the contact between the enhancer and one of its target gene promoters within a TAD, leading to harmful drug events in patients within a drug response cohort. Figure 3A shows how, using various experimental methods, measurements can be obtained in three dimensions from the chromatin spatial interactome and the data analyzed as a two-dimensional plot of enhancer-gene promoter interactions. Figure 3B shows how an SNP can disrupt the chromatin loop between an enhancer and one of two gene promoters regulated within a TAD. This disruption results in the loss of the spatial connection between the enhancer and gene promoter 1, leading to dysregulation of gene 1 and harmful events in this patient and its cohort in response to administration of a specific drug of interest. [Figure 4] A flow diagram and scoring system are shown representing an exemplary method for determining appropriate drugs and dosages for administration to patients suffering from depression or other neuropsychiatric disorders, while avoiding adverse events (AE), based on comparing data from a patient's biological sample to a reference drug-specific pharmacogenomics network and a constituent subnet of the drug of interest. [Figure 5] A flow diagram is shown representing an exemplary method for using the efficacy and adverse event subnetworks obtained from an individual patient's pharmacogenomics network to safely prescribe, proceed with caution, or discontinue prescribing a given drug at a given dosage, or using a quantitative assay of molecular phenotypes generated from the drug's subnetwork. [Figure 6] Two different methods are shown for determining whether a patient of a specific ancestry should take ketamine based on bioinformatics post-analysis and annotation of disease risk SNPs from GWAS, and based on the presence of disease risk SNPs associated with the adverse event subnetwork 3 and efficacy subnetwork 2 of ketamine. [Figure 7]Using machine learning on models constructed from highly heterogeneous hierarchical sets of biomedical and biological data types and elements, we demonstrate how fine-tuning of a set of subnetwork types spanning a range of human drug response phenotypes is achieved, compared to patient input sample data after specific drugs have been selected from a database of pharmacogenomics networks and their constituent subnetworks. [Figure 8] This flowchart illustrates an exemplary method for using similarity scores to match the pharmacodynamic efficacy and adverse events of a patient's drug with those of a reference drug's pharmacogenomics network, enabling the use of TAD matching methods in clinical trials to investigate drug similarity using TAD patterns and / or drug TAD profiles. [Figure 9A] This document presents an exemplary model of how the system integrates heterogeneous hierarchical biomedical and biological data and processes these multiscale data using machine learning and deep learning for pharmacogenetic network topology and subnetwork reconstruction. This strategy for mapping drug networks provides insights into mechanical on-target and off-target effects. The discovered pharmacogenetic network topology provides a foundation for advanced pharmacogenetic decision support, laying the groundwork for subsequent preclinical and clinical research to improve this capability, which can also be used for drug mechanism discovery and drug diversion prediction. [Figure 9B] This flowchart illustrates the method for generating the reconstructed pharmacogenomics network and the corresponding subnetworks for the target drug (including human pharmacogenomics SNP input filters, spatial genomics network reconstruction engines, and iterative gene set optimization engines). Reference sets of information on these pharmacogenomics networks, drug efficacy, and drug adverse events can also be created. [Figure 10] This flowchart illustrates an exemplary method for constructing a pharmacological genomics network and optimizing an integrated drug-gene set for determining its subnetworks. [Figure 11] This diagram illustrates the post-validation of reconstructed pharmacogenomics networks and sets of such networks using bioinformatics data / software and an integrated pharmacological information pipeline. [Figure 12] This example shows a comparison of the significance test results of the intragenetic enhancer SNP rs12967143-G located in the TCF4 gene with other GWAS SNPs, as described, between various neuronal and non-neuronal cell types, using numerical outputs from six different machine learning algorithms used in the analysis. [Figure 13] Figure 13A shows the characteristics of two different ketamine pharmacogenomic subsystems determined from post-hoc validation of ketamine pharmacogenomic networks in the human brain. Figure 13B shows the gene enrichment of ketamine pharmacogenomic subsystems in the human brain that mediate efficacy and neural plasticity. [Figure 14A] This shows pharmacological genomics subnetworks of ketamine in the human brain that mediate efficacy and neuroplasticity, as well as a graphical representation of diseases and conditions associated with efficacy and neuroplasticity. [Figure 14B] This diagram shows the ketamine pharmacological genome subnetworks in the human brain that mediate glutamimate receptor signaling and adverse events, as well as a graphical representation of diseases and conditions associated with this signaling and these adverse events. [Figure 15] List the genes and regulatory RNAs involved in the ketamine efficacy and neural plasticity subnetwork. [Figure 16] This lists the genes involved in ketamine glutamimate receptor signaling and adverse event subnetworks. [Figure 17] This lists the genes involved in the pharmacokinetic and hormone regulatory subnetworks of ketamine. [Figures 18A-18B] Examples of linear regression analysis values ​​for ketamine dose determination and ketamine administration accuracy in the validation cohort are shown. [Figure 19] This document presents definitions of four treatment-resistant depression (TRD) patient subtypes determined by transcranial magnetic stimulation (TMS) combined with neuroimaging of the resting state connection network in the human brain. [Figure 20] This shows examples of neuromaps for four different subtypes of TRD depression patients. [Figure 21] For each of the four distinct subtypes of TRD depression, this document provides recommendations for prescribing / not prescribing medications, as well as examples of alternative medication options. [Figure 22] The combined use of valproic acid and ketamine in the acetylation and deacetylation of H3K9, respectively, leading to neurogenesis and neuronal differentiation, illustrates beneficial combination mechanisms and therapeutic approaches mediated by synergistic histone modifications discovered using the methods described herein. [Figure 23] The complementary pharmacological genomics networks of valproic acid (Figure 23A) and ketamine (Figure 23B) are shown, illustrating neurogenesis and neural differentiation, respectively. [Figure 24] This study demonstrates the combined biological synergistic effects of the pharmacological genomic network of valproic acid and ketamine in neurogenesis, neuronal proliferation, and terminal nerve differentiation. [Modes for carrying out the invention]

[0025] The following text provides a detailed description of numerous different embodiments, but the legal scope of this specification should be understood to be defined by the claims language set out at the end of this disclosure. The detailed descriptions should be interpreted as illustrative only and do not describe all possible embodiments, as it would be impractical, if not impossible, to describe all possible embodiments. Many alternative embodiments may be implemented using either the current art or art developed after the filing date of this patent, but these would still fall within the scope of the claims.

[0026] Unless a term is expressly defined in this Patent by the phrase, "As used herein, the term '______' is defined herein to mean...", or a similar phrase, there is no intention, express or implicit, to limit the meaning of that term beyond its express or ordinary meaning, and it should also be understood that such term should not be interpreted as being limited to the scope based on any description made in any section of this Patent (other than the language of the claims). Where any term used in the last claim of this Patent is referred to in this Patent in a manner that does not contradict a single meaning, this is done solely for clarity to avoid confusing the reader, and such claim term is not intended to be limited by implied or otherwise to its single meaning. Finally, unless an element of a claim is defined by listing the word "means" and function without any detail of any configuration, the scope of any element of a claim is not intended to be interpreted under Section 112, Section 6 of the United States Patent Act.

[0027] This disclosure includes systems and methods for stratifying patients or cohorts of patients diagnosed with a mental disorder or requiring these drugs for other clinical indications for precise drug selection and dosage of NMDAR antagonists. While ketamine is used as an example, these methods can be used with NMDAR antagonists or α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor modulators in clinical trials for the clinical indication of treatment-resistant depression. Ketamine and its enantiomers are novel antidepressants that demonstrate higher efficacy and reduced side effects compared to other antidepressants in patients or specific subsets of patients diagnosed with TRD. The pharmacogenomics decision support system can determine which patients diagnosed with TRD should take ketamine or one of its enantiomers as an antidepressant, and, if so, what the appropriate dosage should be to maximize efficacy, minimize adverse drug events and drug-drug interactions, and reduce the adverse effects of drug-gene and drug-drug interactions for individual patients or patient cohorts. Adverse psychotropic side effects of glutamate receptor targeted drugs for the relief of depression, including psychomimetic and neurocognitive effects, are based on multiple variables, including demographic and sociological variables, trauma history, genotype, and clinical variables, coupled with the heterogeneity of patient populations that may exhibit treatment-resistant depression (TRD).

[0028] This system includes several methods that can be used to construct a clinical decision-making support diagnosis regarding the selection and administration of ketamine or one of its enantiomers as an antidepressant for treatment-resistant depression, and this system can be generalized to other psychiatric drugs. One embodiment includes an integrated multiscale measurement system with components shown in Figure 1D, in which a patient's biological sample can be extensively analyzed as shown in Figure 2, and other relevant clinical data can also be obtained.

[0029] In some embodiments, the system is configured to analyze the minimum amount of results from a biological sample required to match a stored cohort subnetwork reference set. In this case, patient data is compared to a reference set of drug subsystems spanning the entire range of human drug response cohorts using a pre-trained learning machine for the cohort response range of a particular drug. The agreement with the reference set facilitates recommendations for clinical decisions. For example, the reference set of drug subsystems may include a set of reference drug pharmacodynamic efficacy subsystems, reference drug pharmacodynamic adverse event subsystems, reference chromatin remodeling subsystems, and reference pharmacokinetic enzyme and hormone subsystems for a particular drug.

[0030] Electronic encryption brokers are used first to protect health information through anonymization and prevent the identification of patients. Biological samples (or multiple samples) (e.g., blood, cheek swabs, saliva, urine, or other body fluids) are obtained from patients or cohorts of patients accompanied by clinical data from medical records such as electronic health records (EHRs) or other sources. Initial pharmacological and metabolic analysis of small amounts of blood samples or their plasma components collected from patients or cohorts of patients is performed to determine potential drug-drug and drug-gene interactions that may alter subsequent pharmacogenomic decisions. These objective measurements supplement self-reported data, clinician-reported data, or other data contained in EHRs or other patient records.

[0031] Generally speaking, techniques for determining whether or not to administer drugs such as glutamate NMDAR antagonists or partial antagonists, GLRB modulators, or AMPAR agonists to a patient, and / or determining the appropriate dosage of drugs to administer to a patient, can be implemented in a system including one or more client devices, one or more network servers, or a combination of these devices. However, for clarity, the following examples primarily focus on embodiments in which a healthcare professional obtains a patient's biological sample and provides the biological sample to an assay laboratory for analysis.

[0032] Biological samples may include the subject's skin, blood, urine, sweat, lymph, bone marrow, cheek cells, saliva, cell lines, tissues, etc. Cells are then extracted from the biological samples and reprogrammed into stem cells, such as induced pluripotent stem cells (iPSCs). The iPSCs are then differentiated into various tissues, such as neurons and cardiomyocytes, and assayed to obtain the patient's genomic data, chromosomal data, metabolomics data, etc. In some embodiments, the iPSCs may be assayed for gene loci associated with or related to the phenotypic response to the drug of interest. The iPSCs constitute part of a reference set used to derive variables for evaluating individual patients.

[0033] The drug and dosage determination server relies on a drug and dosage determination support engine. This engine receives numerical scores representing the overlap between the input patient sample and the relevant drug-specific reference pharmacogenomics networks stored in a database of 154 such references, as shown in Figure 2. The system uses a pre-trained learning machine that has been trained on the entire range of human drug response cohorts, consisting of a reference set of pharmacogenomics networks for a particular drug. This encompasses the range of human drug response variability to a particular drug, including the reference set that matches the input patient sample.

[0034] The drug and dosage determination server analyzes experimental results to determine the patient's subnetwork representation of a drug gene set, such as ketamine, an NMDAR antagonist or partial antagonist, a GLRB modulator, or an AMPAR agonist. Furthermore, the drug and dosage determination server retrieves, for example, the reference drug's pharmacogenomics network database, the reference pharmacogenomics network and reference constituent subnetwork for the NMDAR antagonist or partial antagonist, GLRB modulator, or AMPAR agonist. Next, the drug and dosage determination server compares the patient's drug subnetwork representation to the reference pharmacogenomics network and reference constituent subnetwork for this drug to determine whether or not to administer this drug to the patient. For example, the drug and dosage determination server may compare the patient's efficacy drug-specific (e.g., ketamine) subnetwork to a reference efficacy drug-specific (e.g., ketamine) subnetwork, or the patient's adverse event drug-specific (e.g., ketamine) subnetwork to a reference adverse event drug-specific (e.g., ketamine) subnetwork. Next, the drug and dosage determination server may decide that the patient should be administered the drug if the similarity between the drug-specific subnetwork for patient efficacy and the drug-specific subnetwork for reference efficacy is greater than a threshold (indicating a high probability that the drug will be effective for the patient). The drug and dosage determination server may also decide that the patient should be administered the drug if the similarity between the drug-specific subnetwork for patient adverse events and the drug-specific subnetwork for reference adverse events is below a threshold (indicating a low probability that the patient will experience adverse events), or based on a combination of the two.

[0035] Therefore, the drug and dosage determination server provides a healthcare professional's client device with a recommendation indicating that the patient should take the medication, thereby allowing the healthcare professional to administer the medication to the patient. As a result, the healthcare professional can administer the medication to the patient. In some embodiments, the drug and dosage determination server may determine the dosage of the medication to be administered to the patient according to a medication algorithm. The medication algorithm may be determined using machine learning techniques such as linear regression and may be based on the patient's demographic data, the patient's clinical data, the patient's biological data, etc.

[0036] The drug and dosage determination server supports regression algorithms (e.g., least squares regression, linear regression, logistic regression, stepwise regression, multivariate adaptive regression splines, local estimation scatter plot smoothing, etc.), instance-based algorithms (e.g., k-nearest neighbors, learned vector quantization, self-organizing maps, locally weighted learning, etc.), regularization algorithms (e.g., ridge regression, least absolute contraction and selection operators, elastic networks, minimum angle regression, etc.), and decision tree algorithms (e.g., classification trees and regression trees, ID3 (iterative dichotomizer), C4).5, C5, Chi-squared automatic interaction detection, decision strains, M5, conditional decision trees, etc.), clustering algorithms (e.g., k-means, k-median, expectation maximization, hierarchical clustering, spectral clustering, mean shift, density-based pharmacogenomic clustering of applications with noise, OPTICS (ordering points to identify the clustering structure), etc.), correlation rule learning algorithms (e.g., a priori algorithm, Eclat algorithm, etc.), Bayesian algorithms (e.g., naive Bayes, Gaussian naive Bayes, polynomial naive Bayes, AODE (averaged one-dependence) estimators), Bayesian trust networks, Bayesian networks, etc., artificial neural networks (e.g., perceptron, Hopfield network, radial basis function network, etc.), deep learning algorithms (e.g., multilayer perceptron, deep Boltzmann machine, deep belief network, convolutional neural network, stacked autoencoder, generative adversarial network, etc.), dimensionality reduction algorithms (e.g., principal component analysis, principal component regression, partial least squares regression, summon mapping, multidimensional scaling, projection tracking, linear discriminant analysis, mixture discriminant analysis, quadratic discriminant analysis, flexible discriminant analysis, factor analysis, independent component analysis, non-negative matrix factor analysis) A variety of machine learning techniques, including but not limited to sub-decomposition and t-distribution stochastic nearest neighbor embeddings, ensemble algorithms (e.g., boosting, bootstrap aggregating, AdaBoost, stack generalization, gradient boosting machines, gradient boosting regression trees, random decision forests, etc.), reinforcement learning (e.g., delayed learning, Q-learning, learning automata, SARSA (State-Action-Reward-State-Action), etc.), support vector machines, mixture models, evolutionary algorithms, and stochastic graphical models, may be used to determine the dosage of drugs to be administered to a patient and to perform other methods described herein.

[0037] Referring to Figure 1A, an exemplary pharmacogenetic decision support system 100 determines whether or not to administer psychotropic drugs to patients suffering from depression, such as patients with TRD, and the appropriate dosage of psychotropic drugs. The pharmacogenetic decision support system 100 includes a drug and dosage decision server 102 and a number of client devices 106-116 that can be communicated via a network 130, as described below. In one embodiment, the drug and dosage decision server 102 and the client devices 106-116 may communicate by radio signals 120 via a communication network 130, which may be any preferred local or wide-area network(s) including a WiFi network, a Bluetooth network, a cellular network such as 3G, 4G, Long-Term Evolution (LTE), 5G, or the Internet. In some examples, the client devices 106-116 may communicate with the communication network 130 via an intermediate radio or wired device 118, which may be a wireless router, a wireless repeater, or a base transceiver base station of a mobile phone service provider. For example, client devices 106-116 may include a tablet computer 106, a smartwatch 107, a network-enabled mobile phone 108, a wearable computing device such as Google Glass® or Fitbit® 109, a personal digital assistant (PDA) 110, a mobile device smartphone 112 (also referred to herein as a “mobile device”), a laptop computer 114, a desktop computer 116, a wearable biosensor, a portable media player (not shown), a phablet, and any device configured for wired or wireless RF (radio frequency) communication. Furthermore, any other suitable client device for recording patient clinical data may also communicate with the drug and dosage determination server 102.

[0038] Each of the client devices 106-116 can interact with the drug and dosage determination server 102 to receive recommendations on whether or not to administer psychotropic drugs to a patient and on the dosage of those drugs. The client devices 106-116 can present these recommendations to healthcare professionals via a user interface.

[0039] In an exemplary implementation, the drug and dosage determination server 102 may be a cloud-based server, application server, web server, etc., and includes memory 150, one or more processors (CPUs) 142 such as microprocessors connected to the memory 150, a network interface unit 144, and an I / O module 148 which may be, for example, a keyboard or touchscreen.

[0040] The drug and dosage determination server 102 may also be communicatively connected to a database 154 of constituent subnetworks, such as a pharmacological genomics network for reference drugs and drug efficacy and adverse events subnetworks.

[0041] Memory 150 may be tangible non-volatile memory and may include any type of suitable memory module, including random access memory (RAM), read-only memory (ROM), flash memory, and other types of persistent memory. For example, memory 150 may store instructions executable by processor 142 for an operating system (OS) 152, which may be any type of suitable operating system, such as a modern smartphone operating system. Memory 150 may also store instructions executable on processor 142 for, for example, a drug and dosage determination support engine 146. The drug and dosage determination server 102 is described in more detail below with reference to Figure 1B. In some embodiments, the drug and dosage determination support engine 146 may be part of one or more of the client devices 106-116, the drug and dosage determination server 102, or a combination of the drug and dosage determination server 102 and client devices 106-116.

[0042] In any case, the drug and dose determination support engine 146 can obtain experimental results from patient biological samples only when necessary to match one of a set of pharmacogenetic networks and their constituent subnetworks that define human drug response variability to a particular drug. These include molecular data, including genomic variability defined by SNPs in PD and PK genes, pharmacogenetic interactions between regulatory elements, genes in the patient's genome that can be defined using chromosomal conformational data such as Hi-C, and / or direct topologically related domain (TAD) specific measurements (including differential gene expression determined using RNA sequencing (RNA-Seq) or expression microarray profiling), and patient-specific TAD contactome measurements in relevant or surrogate cell types using chromosomal conformational capture (e.g., 3C, 4C, 5C, Hi-C, ChIA-PET, and GAM). The molecular data can be assayed for loci associated with or causally associated with the phenotypic response to the drug of interest. Furthermore, the drug and dosage determination support engine 146 can obtain a reference pharmacological genomics network and reference configuration subnetwork for the target psychotropic drug (e.g., ketamine) from the pharmacological genomics network database 154 of reference drugs.

[0043] Next, the drug and dosage determination support engine 146 can analyze the patient's experimental results to determine the subnetwork representation of the target psychotropic drug, including efficacy subnetworks and adverse event subnetworks. The drug and dosage determination support engine 146 can compare the patient's efficacy subnetworks and adverse event subnetworks to reference efficacy and adverse event subnetworks to determine whether or not to administer the target psychotropic drug to the patient. If the similarity between the drug-specific subnetwork for the patient's efficacy and the drug-specific subnetwork for reference efficacy is greater than a threshold, and / or the similarity between the drug-specific subnetwork for the patient's adverse events and the drug-specific subnetwork for reference adverse events is below a threshold, the drug and dosage determination support engine 146 can determine that the patient should be administered the target psychotropic drug. The drug and dosage determination support engine 146 can then provide a recommendation to the healthcare professional client devices 106-116 indicating that the patient should take the target psychotropic drug. Otherwise, the drug and dosage determination support engine 146 may provide a recommendation for another drug to be administered to the patient to treat depression. Furthermore, the drug and dosage determination support engine 146 can determine the dosage of the psychotropic drug to be administered to the patient according to a medication algorithm. The drug and dosage determination support engine 146 can also provide the recommended dosage of the psychotropic drug to the client devices 106-116 of healthcare professionals.

[0044] The drug and dosage determination server 102 can communicate with client devices 106-116 via the network 130. The digital network 130 may be a private network, a secure public internet, a virtual private network, and / or several other types of networks, such as a dedicated access line, a regular telephone line, a satellite link, or a combination thereof. If the digital network 130 includes the internet, data communication may be conducted via the digital network 130 using internet communication protocols.

[0045] Referring here to Figure 1B, the drug and dosage determination server 102 may include a controller 224. The controller 224 may include program memory 226, a microcontroller or microprocessor (MP) 228, random access memory (RAM) 230, and / or input / output (I / O) circuits 234, all of which may be interconnected via an address / data bus 232. In some embodiments, the controller 224 may also include or be communicably connected to a database 239 or other data storage mechanism (e.g., one or more hard disk drives, optical storage drives, solid-state storage devices, etc.). The database 239 may include data such as pharmacogenomics network reference data, drug recommendation display templates, web page templates, and / or web pages, as well as other data necessary to interact with the user via the network 130. The database 239 may include data similar to that of the database 154 described above with reference to Figure 1A.

[0046] Figure 1B represents a single microprocessor 228, but it will be understood that the controller 224 may contain multiple microprocessors 228. Similarly, the memory of the controller 224 may contain multiple RAMs 230 and / or multiple programmable memories 226. Figure 1B represents the I / O circuit 234 as a single block, but the I / O circuit 234 may contain many different types of I / O circuits. The controller 224 may implement the RAM(s) 230 and / or programmable memories 226 as semiconductor memory, magnetically readable memory, and / or optically readable memory.

[0047] As shown in Figure 1B, the program memory 226 and / or RAM 230 may store various applications for execution by the microprocessor 228. For example, a user interface application 236 may provide a user interface to the drug and dosage determination server 102, thereby enabling, for example, a system administrator to configure, troubleshoot, or test various aspects of the server's operation. The server application 238 may operate to receive patient molecular data, analyze the molecular data to determine the patient's subnetworks related to a specific drug of interest, compare the patient's subnetworks to a reference subnetwork for the specific drug of interest, decide to administer the specific drug of interest to the patient based on the comparison, and send recommendations to client devices 106-116 for administering the specific drug of interest to the patient. The server application 238 may be a single module 238, such as the drug and dosage determination support engine 146, or multiple modules 238A, 238B.

[0048] Although the server application 238 is depicted in Figure 1B as including two modules 238A and 238B, the server application 238 may include any number of modules that accomplish tasks related to the implementation of the drug and dosage determination server 102. Furthermore, although only one drug and dosage determination server 102 is depicted in Figure 1B, it will be understood that multiple drug and dosage determination servers 102 may be provided to distribute server load, serve different web pages, etc. Examples of these dosage determination servers 102 include web servers, company-specific servers (e.g., Apple® servers), and servers located in retail networks or private networks.

[0049] Referring here to Figure 1C, the laptop computer 114 (or any client devices 106-116) may include a controller 242, such as a display 240, a communication unit 258, a user input device (not shown), and a drug and dosage determination server 102. Similar to the controller 224, the controller 242 may include program memory 246, a microcontroller or microprocessor (MP) 248 (random access memory (RAM) 250), and / or input / output (I / O) circuits 254, all of which may be interconnected via an address / data bus 252. The program memory 246 may include an operating system 260, data storage 262, multiple software applications 264, and / or multiple software routines 268. The operating system 260 may include, for example, Microsoft Windows®, OS X®, Linux®, Unix®, etc. The data storage 262 may include data such as application data 264 for multiple applications, routine data 268 for multiple routines, and / or other data necessary for interacting with the drug and dosage determination server 102 via the digital network 130. In some embodiments, the controller 242 may also include, or be communicably connected to, other data storage mechanisms located within the laptop computer 114 (e.g., one or more hard disk drives, optical storage drives, solid-state storage devices, etc.).

[0050] The communication unit 258 may communicate with the drug and dosage determination server 102 via any suitable wireless communication protocol network, such as a wireless telephone network (e.g., GSM, CDMA, LTE, etc.), a Wi-Fi network (802.11 standard), a WiMAX network, or a Bluetooth network. User input devices (not shown) may include a “soft” keyboard displayed on the laptop computer 114's display 240, an external hardware keyboard communicating via a wired or wireless connection (e.g., a Bluetooth keyboard), an external mouse, a microphone for receiving voice input, or any other suitable user input device. As stated with respect to the controller 224, Figure 1C depicts only one microprocessor 248, but it should be understood that the controller 242 may include multiple microprocessors 248. Similarly, the memory of the controller 242 may include multiple RAMs 250 and / or multiple program memories 246. Figure 1C depicts the I / O circuit 254 as a single block, but the I / O circuit 254 may include many different types of I / O circuits. The controller 242 may implement, for example, RAM(s) 250 and / or program memory 246 as semiconductor memory, magnetically readable memory and / or optically readable memory.

[0051] One or more processors 248 may be adapted and configured to run any one or more of a plurality of software applications 264 residing in program memory 246 and / or any one or more of a plurality of software routines 268, as well as other software applications. One of the plurality of applications 264 may be a client application 266 that can be implemented as a set of machine-readable instructions for performing various tasks related to receiving information on the laptop computer 114, displaying information on the laptop computer 114, and / or sending information from the laptop computer 114.

[0052] One of the multiple applications 264 may be a native application and / or web browser 270, such as Apple's Safari®, Google Chrome®, Microsoft Internet Explorer®, and Mozilla Firefox®, which can be implemented as a set of machine-readable instructions for receiving, interpreting, and / or displaying webpage information from the drug and dosage determination server 102, while also receiving input from a user, such as a medical professional or researcher. Another application among the multiple applications may include an embedded web browser 276, which can be implemented as a set of machine-readable instructions for receiving, interpreting, and / or displaying webpage information from the drug and dosage determination server 102.

[0053] One of several routines may include a drug recommendation display routine 272 that presents on display 240 a recommendation on whether or not to administer the psychotropic drug of interest to the patient and / or a recommended dose.

[0054] Preferably, the user can implement the companion diagnostic system 100 by launching a client application 266 from a client device such as one of the client devices 106-116 to communicate with the drug and dosage determination server 102. Furthermore, the user can also implement the companion diagnostic system 100 by launching or instantiating any other suitable user interface application (e.g., a native application or a web browser 270, or any other of the multiple software applications 264) to access the drug and dosage determination server 102.

[0055] Figure 1D shows a flowchart representing an exemplary method 160 for determining the drug and dosage to be administered to a patient suffering from depression, based on comparing data from a patient's biological sample with a reference pharmacological genomics network and constituent subnetworks of the drug of interest. Method 160 may be performed by a drug and dosage determination server 102.

[0056] In some embodiments, patient biosamples are analyzed as shown in Figure 1D for personalized therapy to quantify the relative activation of different pathways mediating the mechanism of action of ketamine in the human CNS, as determined using the method described herein. Patient or patient cohort biosamples can be analyzed using pharmacological metabolomics assays to determine drugs or metabolites in the sample that may cause undesirable drug-drug interactions and affect drug efficacy, adverse events, and administration. Measurements of biological samples include: (1) genotyping of pharmacokinetic SNPs, which have been shown to determine metabolic status (poor, subnormal, normal, or ultrafast subtype) in ketamine specimens consisting of mutations in the CYP2B6 gene; (2) targeting of pharmacodynamic SNPs as input to pharmacogenomic networks and subnetwork profiling, which determine both efficacy and adverse events, as analyzed using a pharmacogenomic genome classifier and a pharmacodynamic subnetwork profiling system; (3) direct topologically related domain (TAD) specific measurements (including differential gene expression determined using RNA sequencing (RNA-Seq) or expression microarray profiling), and patient-specific TAD contactome measurements in related or surrogate cell types using chromosomal conformational capture (e.g., 3C, 4C, 5C, Hi-C, ChIA-PET, and GAM); and (4) pharmacological metabolomics analysis.

[0057] The data processing pipeline shown in Figure 1D consists of parallel routes for analyzing biological samples using multiple methods and analysis of available clinical data from the same patient that may be collected from electronic health records (EHRs). This example shows analyses used to determine whether or not a particular patient should be given an N-methyl-D-aspartate receptor (NMDAR) modulator. Adverse events associated with NMDAR modulators such as ketamine can be severe, including severe dissociation, hallucinations, and nightmares. Therefore, the first decision made from these parallel analyses is to prevent the patient from taking NMDAR modulators such as ketamine if the system predicts that the individual will experience moderate to severe adverse events.

[0058] The data processing pipeline shown in Figure 1D may make a “discontinuation” decision regarding the administration of NMDAR modulators such as ketamine based on the patient’s molecular network representation, specific disease risk SNPs the patient may have (as shown in Figure 23), the patient’s treatment-resistant depression phenotype (as shown in Figures 20 and 22), and dose adjustments based on the patient’s drug metabolism profile. If the system determines that the patient may be taking an NMDAR modulator such as ketamine, the dose determination algorithm is initiated.

[0059] Method 160 is used as a clinical decision-supporting diagnostic to determine whether a patient should be prescribed NMDAR modular as an antidepressant (block 162) and to determine the optimal dosage for the patient. In some embodiments, patient biosamples are analyzed for personalized therapy to quantify the relative activation of different pathways mediating the mechanism of action of ketamine in the human CNS, as determined using the method herein (block 164a). Patient or patient cohort biosamples may be analyzed using drug metabolism assays to determine drugs or metabolites in the sample that may cause undesirable drug-drug interactions and affect drug efficacy, adverse events, and administration. The biological sample measurements in Block 166 include: (1) genotyping of pharmacokinetic SNPs, which have been shown to determine metabolic status (poor, subnormal, normal, or ultrafast subtype) in a ketamine model consisting of mutations in the CYP2B6 gene; (2) targeting of pharmacodynamic SNPs as input to pharmacogenomic networks and subnetwork profiling, which determine both efficacy and adverse events, as analyzed using a pharmacogenomic genome classifier and a pharmacodynamic subnetwork profiling system; (3) direct topologically related domain (TAD) specific measurements (including differential gene expression determined using RNA sequencing (RNA-Seq) or expression microarray profiling), and patient-specific TAD contactome measurements in related or surrogate cell types using chromosomal conformation capture (e.g., 3C, 4C, 5C, Hi-C, ChIA-PET, and GAM); and (4) pharmacological metabolomics analysis (Block 164c).

[0060] As described above, biological sample measurements include targeting pharmacodynamic SNPs as input to pharmacogenetic network and subnetwork profiling (blocks 168, 170b). This determines both efficacy and adverse events when analyzed using a pharmacogenetic genome classifier and pharmacodynamic subnetwork profiling system. In block 170a, the reference pharmacogenetic network and subnetworks for the drug of interest are retrieved from a reference database. Next, patient subnetworks for the specific drug of interest, including efficacy and adverse event subnetworks, are compared to the reference pharmacogenetic network and subnetworks for the drug of interest (block 172). To determine similarity with the reference set, two distinct pairs of reference patient metrics are the exact measure of similarity and the output similarity scores for each of the efficacy and adverse event subnetworks. In block 172, the similarity scores of the efficacy and adverse event subnetworks for the drug of interest can be used to decide whether or not to administer the drug of interest to the patient. For example, if the similarity score of the efficacy subnetwork exceeds the threshold similarity score, method 160 can determine that the drug of interest should be administered to the patient (block 176). Otherwise, method 160 decides to select a different drug (block 174).

[0061] In addition to comparing the patient subnetwork for the drug of interest with the reference subnetwork for the drug of interest, block 164b collects and analyzes the patient's clinical data to determine whether or not to administer the drug to the patient and / or the dosage of the drug. More specifically, the patient's HAMD score and / or symptoms can be analyzed to classify the patient into one of four TRD patient subtypes (block 180). The TRD patient subtypes are described in detail below with reference to Figures 15-17. If the patient is classified as TRD subtype 3 (block 182), method 160 determines to select a different drug (block 184). Other clinical data such as the patient's drug-drug interactions, age, weight, biological sex, weight index, ethnicity, family history, patient history of drug abuse, diagnostic code, hospitalization history, drug-gene interactions, psychosis history, and whether the patient smokes or uses nicotine can also be analyzed (block 186).

[0062] Next, the dosage of the target drug to be administered to the patient is determined (block 178). The dosage can be determined based on a medication algorithm that has predetermined constants applied to each of several patient characteristics, such as biological characteristics, demographic characteristics, and clinical characteristics. In other embodiments, the medication algorithm can be generated using machine learning techniques. Patient features utilized in the dosing algorithm may include biological data such as SNPs that have been reported to stratify human responses to ketamine. Patient features may also include demographic data of the patient, such as the patient's sex, height and weight, age, and ethnicity. Furthermore, patient features may include clinical data such as family history, drug-drug interactions, history of mental illness, whether the patient smokes or uses nicotine, and Hamilton Depression Scale (HAM-D) scores.

[0063] Figure 2 illustrates various measures that can be taken from patient biological samples. In one embodiment, blood and cheek swab samples are obtained and processed. For ketamine and other drugs that undergo first-pass metabolism by proteins encoded by the superinducible and supervariable CYP2B6 gene, genotyping of target SNPs is performed using a 4-SNP panel, but most importantly, testing for the splicing variant SNP rs3745274. This is because the splicing variant SNP rs3745274 is relatively common among the human population (>10% frequency), and carriers of this SNP include very low metabolizers of any drug metabolized primarily by this enzyme, and are very likely to experience adverse drug events from NMDAR modulators such as ketamine.

[0064] Based on the method shown in Figure 1D and the measurements collected in Figure 2, it is recommended to perform drug metabolism analysis of the patient's blood to rule out the possibility of negative drug-drug or drug-gene interactions.

[0065] Figure 2 illustrates analyses that can be performed on patient biological samples to obtain the minimum information necessary to match machine or deep learning algorithms to drug-specific patterns in a comprehensive subnetwork reference set containing topologically related domain (TAD) activations included in database 154. These methods include measuring TAD-specific changes in gene expression using RNA-seq or expression microarrays, pharmacogenomic contacts between genes using chromosomal conformation capture analysis such as Hi-C, and / or targeted genotyping after or in anticipation of drug administration to a patient, through analysis of cheek swabs obtained from the patient before deciding whether or not to administer the drug.

[0066] The biological sample measurements in Figure 2 include genotyping of targeted pharmacokinetics and pharmacodynamic SNPs (block 292), direct TAD-specific measurements such as differential gene expression and pharmacogenomic contacts (block 294), and pharmacological metabolomics (block 296). Genotyping of targeted pharmacokinetics and pharmacodynamic SNPs includes identifying mutations in the CYP2B6 gene, which has been shown to determine metabolic status (block 282), and classifying the intensity of efficacy and adverse event subnetworks according to profiling of pharmacogenomic networks and subnetworks (block 284). Direct chromatin contact-specific measurements include differential gene expression determined using circumscribed RNA sequencing (RNA-Seq) (block 286), and circumscribed chromosome conformational capture analysis to identify pharmacogenomic contacts (block 288). Furthermore, pharmacological metabolomics analysis can be used to identify existing drugs and metabolites in a patient's biological sample and to evaluate potential drug-drug interactions, for example (block 290), and can be used to regularly monitor the patient's drug-drug interactions and medication adherence.

[0067] Figure 3 illustrates a simple example of how a SNP within an enhancer in a network can disrupt the contact between the enhancer and one of its target gene promoters within the TAD, potentially causing adverse drug events in patients within a drug response cohort. Figure 3A shows how, using various experimental methods, three-dimensional measurements can be obtained from the chromatin pharmacogenomics interactome, and the data can be analyzed as a two-dimensional plot of enhancer-gene promoter interactions. Figure 3B illustrates how a SNP can disrupt the chromatin loop between the enhancer and one of the two gene promoters it regulates within the TAD. This disruption leads to a loss of pharmacogenomic connectivity between the enhancer and gene promoter 1, resulting in dysregulation of gene 1 and adverse events in this patient and its cohort in response to administration of the specific drug of interest.

[0068] Figure 4 shows the first of several embodiments for matching reference data with patient input for clinical decision support. In this example, a reference set of pharmacological genomic drug subnetworks for a specific drug can be combined with sparse results from patient biosample inputs, and a combinatorial cotraining method can be used to derive drug efficacy scores for decision-making.

[0069] For example, as shown in Figure 5, the drug and dosage determination server 102 compares the patient efficacy and adverse event subnetworks of the psychotropic drug of interest with the reference efficacy and adverse event subnetworks of the psychotropic drug of interest. In this example, reference subnetwork 1 (reference number 502) and reference subnetwork 2 (reference number 504) are adverse event subnetworks, and reference subnetwork 3 (reference number 506) is the efficacy subnetwork. Patient X's subnetwork 510 is different from the adverse event subnetworks (reference subnetworks 1 and 2 (reference numbers 502 and 504)) and similar to the efficacy subnetwork (reference subnetwork 3 (reference number 506)). Therefore, the drug and dosage determination server 102 determines that the psychotropic drug of interest should be administered to patient 520. Patient Y's subnetwork 512 is different from one of the adverse event subnetworks (reference subnetwork 1 (reference number 502)) but similar to the other adverse event subnetworks and the efficacy subnetwork (reference subnetworks 2 and 3 (reference numbers 504 and 506)). Due to an adverse event in subnetwork 2 (reference number 504), the drug and dosage determination server 102 determines that the target psychotropic drug should be administered to the patient, but at a reduced dose 522. Patient Z's subnetwork 514 is similar to both the adverse event subnetworks (reference subnetworks 1 and 2 (reference numbers 502 and 504)) and different from the efficacy subnetwork (reference subnetwork 3 (reference number 506)), so the drug and dosage determination server 102 decides not to administer the target psychotropic drug to patient 524.

[0070] Figure 5 illustrates an example of how bioinformatics analysis can be used to match a drug reference set of antidepressant efficacy versus adverse event signatures obtained from a gene set optimizer with those from an input patient's biological sample. In Figure 5, based on the agreement of antidepressant efficacy from subnetwork 2 of the reference pharmacogenomics network and the decreasing agreement of adverse events to antidepressant efficacy from subnetwork 3 of the pharmacogenomics network that the individual would likely experience, ketamine can be administered to patient X. However, in patient Z, the ketamine dosage needs to be adjusted because this individual has some of the disease risk SNPs from the GWAS found in the subnetworks of the pharmacogenomics network.

[0071] Figure 6 illustrates how two different methods can be derived based on post-hoc bioinformatics analysis or annotation of disease risk SNPs from GWAS to determine whether a patient should not take ketamine, based on the presence of disease risk SNPs associated with the ketamine adverse event subnetwork and efficacy subnetwork of the ketamine pharmacogenomics drug network. These simple and preliminary methods can be used to initially screen patients for potential negative consequences of ketamine administration.

[0072] Biological samples 2002 and 2004 are collected from patients A and B. Biological sample 2002 from patient A is analyzed to perform pharmacokinetic SNP targeting as input to profiling of the pharmacogenetic network and subnetworks and to determine the efficacy and adverse event subnetworks for patient A (Block 2006). Biological sample 2004 from patient B is analyzed to identify pharmacokinetic SNPs associated with the ketamine response (Block 2008). Next, the efficacy and adverse event subnetworks for patient A are compared to the reference pharmacogenetic network and reference efficacy and adverse event subnetworks for ketamine (Block 2010). SNPs from patient B are compared to SNPs included in the reference pharmacogenetic network and the reference efficacy and adverse event subnetworks for ketamine (Block 2012). In Figure 6, before administering ketamine, the dose for patient A should be adjusted based on the agreement between the ketamine subnetwork mediating adverse events and the antidepressant efficacy derived from the pharmacogenetic network the individual is likely to experience, as well as the decrease in agreement between the antidepressant efficacy subnetwork of the reference pharmacogenetic network (Block 2014). Similarly, in patient B, the dose of ketamine administered to patient B should be adjusted because this individual has many of the disease risk SNPs from the GWAS found in the subnetwork of the pharmacogenetic network (Block 2016).

[0073] Figure 7 illustrates a comprehensive strategy for determining whether ketamine should be administered to a particular patient and how other clinical data from that patient can be fine-tuned to make informed clinical decisions, using a library of human drug response cohort phenotypes stored in a reference database.

[0074] Figure 8 shows a flowchart illustrating how a well-known pattern matching algorithm from deep learning in computer vision is used to match the efficacy and adverse events of a patient's drug with those of a reference drug's pharmacological genomics network using similarity scores. Figure 8 also shows how this strategy can be used to discover novel similar drugs.

[0075] Figure 9A provides an overview of the integrated multiscale data analysis used in the system. Figure 9B shows a flowchart illustrating an exemplary method for generating the reconstructed pharmacogenomics network and the corresponding subnetworks for the drug of interest (e.g., human pharmacogenomics SNP input filters, pharmacogenomics network reconstruction engine, and iterative gene set optimization engine).

[0076] As described above, biological sample measurement includes targeting of pharmacodynamic SNPs as input to pharmacogenetic network and subnetwork profiling. This determines both efficacy and adverse events when analyzed using a pharmacogenetic genome classifier and a pharmacodynamic subnetwork profiling system. Next, the patient subnetwork for the specific drug of interest is compared to a reference pharmacogenetic network and its constituent subnetworks for the drug of interest. Figure 9B shows a method 300 for identifying a reference pharmacogenetic network and constituent subnetworks for a specific drug of interest, such as ketamine. In some embodiments, the drug and dosage determination server 102 performs method 300 to identify the pharmacogenetic network and constituent subnetworks for the specific drug of interest and stores the pharmacogenetic network and constituent subnetworks in the reference pharmacogenetic network database 154. In other embodiments, method 300 is performed by a separate computing device, and the output of this method is provided to the drug and dosage determination server 102 and stored in the reference pharmacogenetic network database 154.

[0077] SNP selection In any case, in block 302, SNPs are obtained from human clinical studies that have shown a significant association with the response to and adverse events of the drug of interest. Because the location of SNPs related to the trait under study is, in most cases, inaccurately assigned to the nearest gene or a close candidate gene in linear sequence-by-sequence GWAS of the published literature and reference human genome assembly, accurate localization and annotation techniques using permutation are used to determine the actual location of the reported SNP.

[0078] The new research has several important implications for identifying drug pharmacogenetic networks. First, target mechanisms of novel drugs can be identified by collecting pharmacogenetic network outputs in a training set using computer vision-based TAD matching with deep learning (machine learning) and validation using correspondences to known drug-induced genome-wide TAD matrices. Second, clustering of novel drug target mechanisms in previously defined but incompletely informed biological pathways will increase the likelihood of success. Third, insights gained using three-dimensional (3D) genome architectures to determine drug targets from pharmacogenetic GWAS will lead to next-generation drug candidates and significantly improve the accuracy of pharmacogenetic clinical decision-making support diagnostics.

[0079] In Block 304, pharmacodynamics, pharmacokinetics, and other SNPs are evaluated using the pharmacogenomics informatics pipeline. The pipeline uses read SNPs reported from GWAS and candidate gene studies to find genetically linked tolerable candidate SNPs using TAD boundaries instead of linkage disequilibrium measurements. These SNPs are evaluated in two distinct workflows: an enhancer-modulation workflow for regulatory SNPs and a coding SNP workflow. The enhancer-modulation SNP workflow evaluates tolerable candidate SNPs in disease-associated tissues for DNA methylation, transcription factor binding, histone marks, DNase I hypersensitivity, chromatin state, quantitative trait loci (QTLs), and transcription factor binding site disruption using tissue-specific omics datasets. The coding SNP workflow examines histone modifications to find common non-synonymous coding SNPs within the pool of tolerable candidate SNPs and exclude enhancer SNPs containing exons. Both sets of SNPs are mapped to their TADs and host genes and screened for expression in relevant tissues. Next, open-source machine learning algorithms are used to evaluate the final output SNPs to determine whether they are causative (block 306), and causative variants are retained for further analysis in the workflow (block 308). Exonic SNPs are also evaluated as splice donors or splice acceptors using the Altrans algorithm. If exonic SNPs are found to be involved in alternative splicing, they are stored as such.

[0080] Using a casual enhancer SNP for the question In Block 310, enhancer SNPs are used as probes to determine target genes within the same TAD where the enhancer is located, and pharmacogenomic interactions with other TADs are determined using Hi-C chromosome conformation capture and ChIA-PET datasets generated from cell types and tissues on which the drug of interest acts (Block 314). Genes containing other functional elements, such as long non-coding RNAs, are selected for pharmacogenomic networks if they are located within the same TAD that is a target of an enhancer that significantly alters the drug response in the human population, and if the TAD has a strong boundary as predicted by the amount of bound CTCF and a significant association with the super-enhancer (Block 312). Next, the top three statistically significant pharmacogenomic contacts constituting the first set of pharmacogenomic TADs within the same cell and / or tissue type on which the drug of interest acts are evaluated, and genes within these “trans-TADs” are selected if they are controlled by the same cell and / or tissue-specific enhancers on which the drug of interest acts (block 316).

[0081] In block 318, combined sets of genes are evaluated for interconnections, with genes selected from the initial set of TADs containing pharmacogenetic SNPs, and genes selected from "trans-TADs" that contain genes regulated in coordination with the initial set of TAD genes. For example, third-party software such as Ingenuity Pathway Analysis™ can be used to examine the connectivity of combined sets of genes. Using Fisher's right-hand exact test, if significant interconnections exist within the combined sets of genes based on published literature, the genes are placed within a reserve set of genes that constitute the pharmacogenetic network of the drug of interest. Genes that do not form an interconnected network are discarded as non-candidate genes for the pharmacogenetic network (block 320).

[0082] Revision of the Knowledge Base for the Preliminary Pharmacological Genomics Network of Drug-Specific Interconnected Genes Next, in block 322, manual, semi-automated, or automated curation, or a combination thereof, is performed on each gene in this gene set, which includes the preliminary pharmacogenomics network, and if it is determined that the function of a gene is particularly affected by the drug of interest in the cells and / or tissue types it acts on, then the function of that gene is either removed if it is not relevant to the drug of interest in the cells and / or tissue types it acts on, or other genes that are not part of this preliminary set of pharmacogenomics network should be added to the set. The questioning steps include defining the function of the individual gene, the phenotypic consequences of the impairment caused by the gene mutation, and the human cells and tissues in which the gene is expressed, and determining whether it can be a candidate for membership in the pharmacogenomics network of the specific drug of interest.

[0083] In one embodiment, these decisions can be made using manual, semi-automated, or automated strategies, combining the curation of each gene, its mutation profile, and the localization of its expression within human tissues. These are achieved through various web-based search tools, such as gene definitions, genome browser annotations, GWAS catalogs, and other bioinformatics resources. For example, application programming interfaces (APIs) may include executable files written in R, Python, Perl, or other programming languages ​​to facilitate data access, data cleansing, and data analysis. While this embodiment is an enhanced model of manual curation, it can be time-limited, especially when many genes exist within a gene set or subset of genes in a pharmacogenomics network, particularly when functional genomic elements may include functional RNAs such as regulatory RNA or long non-coding RNAs, or when the function of the gene is not well understood. Because these databases are the most comprehensive, enumerating and analyzing the mutation status of a given gene (±10Kb upstream and downstream) is the simplest of the three question steps to perform. Other resources exist for analyzing the tissue distribution of gene expression patterns. When these patterns are compared to the sites on which a specific drug of interest acts, the results of imaging modalities can be analyzed, including radiological studies, optical microscopy in pathology, and more advanced methods. In some embodiments, this analysis is performed using machine learning techniques such as neural networks.

[0084] In another embodiment, a Bayesian probabilistic classifier can be used based on machine learning or using Bayesian probabilistic computation. The complexity of data analyzed from heterogeneous data resources can be reduced using automated methods, in which functional knowledge profiles of genes, their mutation status, and their tissue expression mappings are inputs to a learning machine that is trained on many such instances and individually tested on different sets of instances to determine accuracy. The predictive functions selected by the trained neural network can be implemented in an aided vector machine classifier to build a gene function and mutation prediction model. The subsequent machine state determines the validity of the statistical fit to the pharmacogenomics network.

[0085] In some scenarios, machine learning can suffer from overfitting, resulting in false positives or false negatives. In another embodiment, a semi-automated simple Bayesian classification can be performed using machine learning in parallel to improve the accuracy of the final output.

[0086] Knowledge base curation can be performed in the following steps: First, gene definitions are examined from multiple databases to understand whether they are specifically, rather than comprehensively, affected by the drug of interest. Furthermore, published literature, including gene names or precursor gene names or equivalent protein names, as well as text word strings containing functions related to the drug of interest, is evaluated according to a thorough internet search using, for example, Google Scholar® and / or PubMed. These may include binding affinity studies that reproducibly find molecules that bind with an affinity of no more than 10 times the affinity that the drug of interest binds to the same pharmacodynamic target. Second, the drug and dose determination server 102 examines each gene for all mutations, including SNPs, a variable number of tandem repeats, duplications, and all other known variant changes, extending linearly within ±10kb from the transcription start site(s) and stop codon(s) of the gene examined in a genome browser such as the UCSC genome browser or the Ensembl genome browser. If any of these mutations are found in any source, such as published literature or unpublished clinical trial data, and they are involved in the action of the drug of interest, such as efficacy, adverse events, or first-pass metabolism, they are paired with a preliminary set of genes that make up the pharmacogenomics network (block 324). Thirdly, especially in complex tissues such as the brain, skin, and cardiovascular system, the drug and dosage determination server 102 performs a qualitative matching mapping to compare the expression of all genes in this final set with the location (if known) where the drug of interest exerts its action. Genes whose expression does not match the pharmacodynamic substrate of the drug of interest are discarded (block 324). Finally, the connections of this gene set are examined using third-party software such as Ingenuity Pathway Analysis® (block 326). If the drug and dosage determination server 102 determines, using Fisher's right-hand exact test, that significant interconnections exist based on published literature, they are placed in the preliminary set of genes that make up the pharmacogenomics network of the drug of interest.Genes that do not form a connected network are discarded as non-candidate genes in the pharmacogenomics network (block 328).

[0087] Optimization of repeating gene sets As shown in block 330 and in more detail in Figure 10, the optimization of a repeating gene set is performed on a set of candidate genes identified in the pharmacogenomics network of a specific drug of interest. An exemplary method 400 for the optimization of a repeating gene set to decompose the pharmacogenomics network into subnetworks is shown in the flowchart of Figure 10. The optimization of a repeating gene set can be performed to identify subnetworks in the pharmacogenomics network. More specifically, the optimization of a repeating gene set involves converting all input molecular terms to gene or long non-coding RNA names from, for example, Human Gene Nomenclature Committee (HGNC) names using an API (block 402). The optimization of a repeating gene set differs from gene set enrichment methods in that it combines various statistical methods and does not hierarchically rank genes like threshold-dependent methods. The optimization of a repeating gene set does not rely on the comparison of experimental results such as whole-distribution studies. Instead, the optimization of the repeating gene set uses the Jaccard distance to group genes or long non-coding RNAs from a pharmacogenomics network (block 404) and first measures the similarity between two genes or long non-coding RNAs based on the dissimilarity of a user-selected term. Here, the Jaccard distance is expressed as the ratio of the sizes of symmetric difference gene AΔ gene B=A∩BA∪B to the union (block 406). This can be extended to clusters of related different gene names. The drug and dosage determination server 102 then automatically sorts these sets into subsets of clustered subsets of functionally related genes using a minimum entropy sort algorithm such as the COOLCAT algorithm (block 408), or using a user-defined number of clusters. Following the optimization of gene subsets using entropy minimization, manual curation can be used to assign efficacy, adverse events, or functional mechanism subnetworks based on known attributes of the mechanism of action of the drug under consideration (blocks 410, 412).

[0088] Post-hoc validation using third-party bioinformatics tools For the scientific validation of the decomposition of pharmacogenomics networks into mechanical subnetworks based on the optimization of functional gene subsets, each subnetwork of pharmacogenomics networks will be retrospectively evaluated for, for example, top-level Gene Ontology terms (molecular function and biological processes) determined using other proprietary or open-source pathway analysis software, top-level standard pathways, disease risk gene variant analysis determined using other proprietary or open-source pathway analysis software, and determination of upstream heterologous regulators using different bioinformatics resources (Block 332). Furthermore, the GWAS catalogs of the European Bioinformatics Institute, the National Human Genome Research Institute, and the National Institutes of Health can be searched to find significant SNP characteristics of each gene in each subnetwork's gene set. By providing examples of statistically significant SNPs from GWAS, additional evidence may be provided that mutational disorders in genes contained within each subnetwork offer insights into the normal and unimpaired function of the subnetwork.

[0089] In some embodiments, after post-validation is performed as shown in Figure 11, the resulting pharmacogenetic network and constituent subnetworks of the drug of a particular interest are stored in a database 154, for example, as shown in Figure 1A.

[0090] For example, to map SNPs that cause discretization of ketamine responses in the human population, their target genes within TADs, and the pharmacological-genomic contacts of these TADs, Hi-C chromosome conformational capture data are used in samples obtained from A735 astrocyte cell lines, H1 neuron cell lines, SK-N-SH cell lines, and postmortem human brains.

[0091] Figure 12 shows comparative results of eight different algorithms testing the predicted causal relationship of the GWAS SNP rs12967143-G, an intragenetic enhancer located in the TCF4 gene, a member of the ketamine pharmacogenetic efficacy subnetwork, to other GWAS SNPs, as described using numerical outputs from the machine learning algorithms used in the analysis (*p≦0.05;**p≦0.01;ANOVA).

[0092] Using H-GREEN, a user-adjustable binning software package, overlapping trans-TADs are mapped across different data sources. The top three genome-wide trans-TAD pharmacogenomic contacts can be selected for each original causative SNP TAD locus. Using prior knowledge of ketamine as an anesthetic and analgesic, the top trans-TAD contacts can be scored using the method described herein in recent studies and clinical trials of ketamine and other glutamimate receptor modulators as antidepressants, and for each causative SNP, the top three can be selected for inclusion in the pharmacogenomic network. The recent availability of databases of validated enhancers and their targets can be used for both the original TADs and target TADs in this workflow to reconstruct the pharmacogenomic network for ketamine.

[0093] The intra-TAD and trans-TAD gene sets may serve as seeds to initiate pathway analysis. Filters and thresholds can be applied to exclude genes expressed in cell types, neurons, and astrocytes, as well as in brain regions where ketamine exerts its mechanism of action. These include not only PD genes but also PK genes. PK genes have recently been shown to be highly expressed in the relevant human brain regions where ketamine acts, and in the case of the CYP2B6 gene, this psychotropic drug induces expression at much higher levels than in the liver, gastrointestinal tract, or kidneys.

[0094] Following the output of automated pathway analysis, the validity of the pharmacogenomics network gene set is evaluated, and genes not selected by the pathway analysis program may be added to the pathway. Previous studies of binding affinity using molecular pharmacology methods may add genes whose products exhibit a 10-fold affinity to racemic R, S-ketamine, or enantiomer NMDARs, demonstrating molecular interconnection. Other expression studies and research on ketamine metabolism may yield additional genes to be added to the ketamine pharmacogenomics network.

[0095] The ketamine pharmacogenomics network is analyzed into three subnetworks by optimizing the gene set, two of which are significantly different subsets of genes and regulatory RNAs using repeated analyses. The three subnetworks include (1) antidepressant efficacy and neuroplasticity, (2) glutamimate receptor signaling, chromatin remodeling, and adverse events, and (3) drug-related pharmacokinetics and hormonal regulation. The second subnetwork, (2) glutamimate receptor signaling, chromatin remodeling, and adverse events, may include two distinct subnetworks: a chromatin remodeling subnetwork and a drug pharmacodynamic adverse events subnetwork. Four additional analyses are performed to understand and validate the pharmacogenomics network and its mechanical subnetworks. First, the genes in the pharmacogenomics network and each subnetwork are questioned for the presence of enhancer SNPs related to relevant traits in the GWAS. Second, pathway enrichment, including biological processes and molecular functions, is performed using Gene Ontology standards to determine the most important higher-level pathways of these gene sets. Thirdly, disease gene risk variant analysis is performed. This involves analyzing supersets and subsets of each gene for the overall significance of the contribution of mutations in these sets in humans, in order to appropriately assign them to both the parent pathway and its constituent subnetworks, and assigning the top diseases to the superset and subnetwork sets (most importantly, Fisher's exact test). Fourthly, the top (most importantly, Fisher's exact test) xenobiotic drugs that modulate the supersets of genes constituting the pharmacogenetic network are determined. In the last case, the pharmacogenetic network set of genes needs to be modulated by drugs that mediate the mechanisms of the pharmacogenetic network, but in some subnetworks, depending on the mechanistic attributes of that network, drugs that are more relevant to the subnetwork of the pharmacogenetic network mechanism may be the most significantly relevant.

[0096] Figure 13 shows the higher-level Gene Ontology terminology for two significantly different ketamine pharmacogenomic subnetworks in the human brain. More specifically, Figure 13A shows subnetworks mediating efficacy and neuroplasticity. Figure 13B shows ketamine subnetworks mediating glutamimate receptor signaling and adverse events in the human CNS.

[0097] Figure 14A shows a graphical representation of the pharmacogenomic subnetworks of ketamine in the human brain that mediate efficacy and neuroplasticity. Figure 14B shows a graphical representation of the pharmacogenomic subnetworks of ketamine in the human brain that mediate glutamimate receptor signaling and adverse events in the human brain. More specifically, as shown in Figures 14A and 15, the genes and regulatory RNAs located within the ketamine efficacy and neural plasticity subnetwork include one or more of the following: the Activation Regulatory Cytoskeleton-Associated Protein (ARC) gene, the Achaete-Scute family bHLH transcription factor 1 (ASCL1) gene, the Brain-Derived Neurotrophic Factor (BDNF) gene, the BDNF Antisense RNA (BDNF-AS) gene, the Calcium / Calmodulin-Dependent Protein Kinase II Alpha (CAMK2A) gene, the Cyclin-Dependent Kinase Inhibitor 1A (CDKN1A) gene, the cAMP Response Element Modulator (CREM) gene, the Cut-like Homeobox 2 (CUX2) gene, the DCC Netrin 1 Receptor (DCC) gene, the Dopamine Receptor D2 (DRD2) gene, the Eukaryotic Translation Elongation Factor 2 Kinase (EEF2K) gene, the Fragile X Mental Retardation 1 (FMR1) gene, and the Ganglioside-Induced Differentiation-Associated Protein 1-like 1( GDAP1L1 gene, glutamate receptor 5 (GRM5) gene, Homer scaffold protein 1 (HOMER1) gene, 5-hydroxytryptamine receptor 1B (HTR1B) gene, 5-hydroxytryptamine receptor 2A (HTR2A) gene, Kruppel-like factor 6 (KLF6) gene, Lin-7 homolog C, long non-coding RNA of crumbs cell polarity complex component (LIN7C), long non-coding RNA of LOC105379109, Myocyte enhancer factor 2D (MEF2D) gene, myosin VI (MYO6) gene, myelin transcription factor 1-like (MYT1L) gene, neuronal differentiation 1 (NEUROD1) gene, neuronal differentiation 2 (NEUROD2) gene, neon helix-loop-helix 2 (NHLH2) gene, neuromedin B (NMB) gene, NMDA receptor synaptic nuclear signaling and neuronal migration factor (NSMF) gene, neurotrophic receptor tyrosine kinase 2 (NTRK2) gene,Phosphotase and tensin homolog (PTEN) gene, prostaglandin-endoperoxide synthase 2 (PTGS2) gene, Rac family small molecule GTPase 1 (RAC1) gene, Ras protein-specific guanine nucleotide release factor 2 (RASGRF2) gene, Ras homolog family member A (RHOA) gene, roundabout guidance receptor 2 (ROBO2) gene, long non-coding RNA of RP11_360A181, semaphorin 3A (SEMA3A) gene, SH3 and multiple ankyrin repeat domain 1 (SHANK1) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) gene, SH3 and multiple ankyrin repeat domain 3 (SHANK3) gene, solute carrier family 22 member 15 (SLC22A15) gene, solute carrier family 6 member 2 (SLC6A2) gene, slit guidance ligand 1 (SLIT1) gene, slit guidance ligand 2 (SLIT2) gene, synaptosome-related protein 25 (SNAP25) gene, synapsin I (SYN1) gene, synapsin II (SYN2) gene, synapsin III (SYN3) gene, T-box, brain 1 (TBR1) gene, or transcription factor 4 (TCF4) gene.

[0098] Furthermore, as shown in Figures 14B and 16, the genes and regulatory RNAs located within the ketaming glutamimate receptor signaling and adverse events subnetwork include one or more of the following: acetylcholinesterase (ACHE) gene, activator-7 interacting protein (ATF7IP) gene, activator-7 interacting protein 2 (ATF7IP2) gene, and ATPase The genes for Na+ / K+ transport subunit alpha-1 (ATP1A1), BLOC-1-related complex unit 7 (BORCS7), bromodomain-containing 4 (BRD4), calcium voltage-gated channel subunit alpha-1C (CACNA1C), calcium voltage-gated channel co-subunit beta-1 (CACNB1), calcium voltage-gated channel co-subunit beta-2 (CACNB2), calcium voltage-gated channel co-subunit gamma-2 (CACNG2), cholinergic receptor muscarinic 2 (CHRM2), cholinergic receptor nicotinic alpha-3 subunit (CHRNA3), cholinergic receptor nicotinic alpha-5 subunit (CHRNA5), cholinergic receptor nicotinic alpha-7 subunit (CHRNA7), cannabinoid receptor 1 (CNR1), and disclar djih Molog 3 (DLG3) gene, disclarge homolog 4 (DLG4) gene, DNA methyltransferase 1 (DNMT1) gene, euchromatic histone lysine methyltransferase 1 (EHMT1) gene, gamma-aminobutyric acid type A receptor alpha 2 subunit (GABRA2) gene, gamma-aminobutyric acid type A receptor alpha 5 subunit (GABRA5) gene, glutamate decarboxylase 1 (GAD1) ​​gene, glycine receptor alpha 1 (GLRA1) gene, glycine receptor alpha 2 (GLRA2) gene, glycine receptor beta (GLRB) gene, glutamate ion channel embedded receptor AMPA subunit 1 (GRIA1) gene, glutamate ion channel embedded receptor AMPA subunit 2 (GRIA2) gene, glutamate ion channel embedded receptor AMPA subunit 2 (GRIA4) gene,Glutamate ion channel embedded receptor NMDA subunit 1 (GRIN1) gene, glutamate ion channel embedded receptor NMDA subunit 2A (GRIN2A) gene, glutamate ion channel embedded receptor NMDA subunit 2B (GRIN2B) gene, glutamate ion channel embedded receptor NMDA subunit 2C (GRIN2C) gene, glutamate ion channel embedded receptor NMDA subunit 2D (GRIN2D) gene, glutamate ion channel embedded receptor NMDA subunit 3A (GRIN3A) gene, glutamate ion channel embedded receptor NMDA subunit 3B (GRIN3B) gene, hyperpolarization-activated cyclic nucleotide-gate potassium channel 1 (HCN1) gene, histone deacetylase 5 (HDAC5) gene, methyl-CpG binding domain Protein 1 (MBD1) gene, M-phase phosphorylated protein 8 (MPHOSPH8) gene, neuronal cell adhesion molecule 1 (NCAM1) gene, nitrate synthase 1 (NOS1) gene, nitrate synthase 2 (NOS2) gene, nitrate synthase 3 (NOS3) gene, NAD(P)H quinone dehydrogenase 1 (NQO1) gene, opioid receptor kappa 1 (OPRK1) gene, opioid receptor mu 1 (OPRM1) gene, roundabout guidance receptor 2 (ROBO2) gene, SET domain branching 1 (SETDB1) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) gene, sigma nonopioid intracellular receptor 1 (SIGMAR1) gene, solute carrier family 6 member 9 (SLC6A9) gene, transcriptional activation suppressor (TASOR) gene, axon microtubule TOG array regulator 2 (TOG The array regulator of axonemal microtubules 2) (TOGORAM2) gene, the triplicate motif-containing 28 (TRIM28) gene, or the zinc finger protein 274 (ZNF274) gene.

[0099] As shown in Figure 17, the genes and regulatory RNAs in the pharmacokinetic enzyme and hormone subnetworks include one or more of the following: late-stage acceleration complex subunit 2 (ANAPC2) gene, cytochrome P450 family 2 subfamily A member 6 (CYP2A6) gene, cytochrome P450 family 2 subfamily B member 6 (CYP2B6) gene, cytochrome P450 family 3 subfamily A member 4 (CYP3A4) gene, disc large homolog 4 (DLG4), eukaryotic elongation factor 2 kinase (EEF2K) gene, and estrogen receptor 1 (ESR1) gene. , the gene for glutamate ion channel-embedded receptor AMPA subunit 1 (GRIA1), the gene for glutamate ion channel-embedded receptor AMPA subunit 2 (GRIA4), the gene for glutamate ion channel-embedded receptor NMDA subunit 1 (GRIN1), the gene for glutamate ion channel-embedded receptor NMDA subunit 2B (GRIN2B), the gene for myosin VI (MYO6), the gene for roundabout guidance receptor 2 (ROBO2), the gene for SH3 and multiple ankyrin repeat domains 2 (SHANK2), or the gene for transcription elongation regulator 1 (TCERG1).

[0100] Automated optimization of repeating gene sets into subnetworks of the psychotropic drug pharmacogenomics network may limit users to investigating other functions of the pharmacogenomics network. As previously mentioned with reference to Figure 10, the ketamine pharmacogenomics network is repeatedly degraded until certain specific genes are no longer significantly associated with a subnetwork or are associated with several subnetworks that have been degraded from the pharmacogenomics network. The gene ESR1 encodes the nuclear hormone receptor for estrogen and regulates the expression of several genes in the ketamine pharmacogenomics network. It is well known that estrogen greatly induces the CYP2B6 gene in both the human brain and other locations.

[0101] The learning architecture for training the pattern matching subnetwork includes pre-training of a reference set (reference number 710). More specifically, in block 704, the drug and dosage determination server 102 develops a patient pattern matching subnetwork derived from a patient input biological sample, along with different pre-trained pattern metrics (block 712) and joint functional representation metrics, including features of the efficacy and adverse event subnetworks. To determine similarity to the reference set (blocks 706, 708), two different pairs of reference patient metrics are provided: an exact measure of similarity, as well as output similarity scores for each of the efficacy and adverse event subnetworks (blocks 714, 716). In block 702, a biological sample obtained from a patient (cheek swab, saliva, blood, or urine sample) undergoes genotyping of target enhancer SNPs and a combination of chromosome conformation capture and RNA-seq. Then, in block 704, the drug and dosage determination server 102 performs the necessary analyses to construct input patient-specific maps of the efficacy and adverse event subnetworks for the specific drug of interest. These patient-specific drug-induced subnetwork patterns can be further processed using Bayesian probability calculations to fill in sparse or missing data. As new patients are input, a pre-trained reference set of drug-specific efficacy and adverse event subnetworks for pattern matching is re-optimized for subsequent patients, generating more accurate measurements of inter-human pharmacogenomic variability with enhanced clinical utility. This matching task assumes that patches pass through the same functional encoding before calculating and outputting similarity scores, significantly improving efficiency while reducing computational requirements.

[0102] Therefore, each input set (reference set (reference number 710) and patient set (reference number 720)) is constructed differently using Bayesian-based probabilistic computing, involving feature set extraction and sparse data inference, thereby improving the accuracy of the reference and patient maps. The trained feature network is based on a "sham" network approach, but with the constraint that the two sets must share the same parameters. Once completed, the trained pattern network of drug-induced patient behavior is combined with the network obtained from the reference database, resulting in pairs of efficacy feature sets and pairs of adverse event feature sets. These provide the basis for developing trained efficacy and trained adverse event metrics, attempting to match all features from the patients with the reference set of the drug of interest. These pairwise matching scores result in separate efficacy and adverse event similarity scores between the reference and the patients.

[0103] In a further embodiment, a reference pattern matching set can be developed for each patient, which can be used to create a patient-specific database of such reference maps, and can be periodically updated as additional biological samples are acquired longitudinally from the patient and obtained over time in the clinical setting or outpatient pharmacy.

[0104] In any case, the drug and dosage determination server 102 can then determine whether or not to administer the psychotropic drug of interest to the patient using the efficacy and adverse event subnetwork similarity scores of the psychotropic drug of interest generated via method 700. For example, the efficacy subnetwork similarity score may be compared to a threshold similarity score. If the efficacy subnetwork similarity score exceeds the threshold similarity score, the drug and dosage determination server 102 can determine that the psychotropic drug of interest should be administered to the patient. The adverse event subnetwork similarity score may also be compared to a threshold similarity score. If the adverse event subnetwork similarity score falls below the threshold similarity score, the drug and dosage determination server 102 can determine that the psychotropic drug of interest should be administered to the patient. In another embodiment, the similarity scores of the efficacy subnetwork and the adverse event subnetwork can be combined or aggregated in any suitable manner. For example, the similarity score of the adverse event subnetwork can be subtracted from the similarity score of the efficacy subnetwork. If the combined score exceeds the threshold similarity score, the drug and dosage determination server 102 can determine that the psychotropic drug of interest should be administered to the patient.

[0105] Otherwise, the drug and dosage determination server 102 may decide not to administer the psychotropic drug of interest and provide recommendations to the healthcare professional client devices 106-116 to administer a different drug to treat the patient's depression.

[0106] In some embodiments, the drug and dosage determination server 102 can determine whether to administer the psychotropic drug of interest to a patient by generating a machine learning model based on training data from drug responses of patients previously prescribed the psychotropic drug of interest. The machine learning model can be generated based on several features of the previous patient. These features include: drug-induced subnetworks of the previous patient, SNPs of PD and PK stratifying the patient by drug response, neuroimaging data, direct TAD-specific measurements including differential gene expression, and clinical variables of the previous patient such as age, weight, biological sex, weight index, ethnicity, family history, patient history of substance abuse, diagnostic codes, hospitalization history, drug-drug interactions, psychotic history, whether the patient smokes or uses nicotine, as well as Hamilton Depression Scale (HAM-D) scores. The drug and dosage determination server 102 can obtain the same features of the current patient, including molecular and clinical data, and apply the features of the current patient to the generated machine learning model to determine whether to administer the psychotropic drug of interest to the patient.

[0107] In addition to deciding whether or not to administer the target psychotropic drug to the patient, the drug and dose determination server 102 determines the dose to be administered to the patient. Figures 18A and 18B show analyses performed on independent cohort datasets to develop a regression model for ketamine dose estimation. Published literature and other sources provide both pharmacokinetic SNP data and clinical values ​​that are helpful in determining doses based on CYP2B6 SNPs and clinical data. As can be seen from the difference analysis, the greatest contributions to ketamine dose are the phenotype rs3745274 of low metabolites, the presence or absence of variants causing exon skipping and loss of highly inducible CYP2B6 first-pass metabolism of ketamine. The quantitative intranasal ketamine dose range in this model derivation cohort was 0.4–0.8 mg / kg. The drug and dose determination server 102 can generate models based on published ketamine clinical trial results obtained from clinicaltrials.org. A common example of a sum based on regression variables that can be included in the ketamine dosing algorithm is shown in the following equation. Dosage = exp[2.00 × rs3745274 + 0.25 × Female (biological sex) + 0.22 × rs3786547 + 0.22 × Clopidogrel + 0.19 × rs11083595 + 0.11 × BSA + 0.20 × Smoking + 0.17 × History of suicide attempts + 0.07 × Age in 10-year increments + 0.09 × Non-Hispanic Caucasian ethnicity + 0.20 × Ticlopidine + 0.15 × Previous psychiatric hospitalizations]

[0108] More specifically, the drug and dosage determination server 102 generates a medication algorithm based on published literature and may have predetermined constants to apply to each of several patient characteristics, such as biological characteristics, demographic characteristics, and clinical characteristics, as shown in the above formula. In other embodiments, the drug and dosage determination server 102 can generate a medication algorithm using machine learning techniques. For example, the drug and dosage determination server 102 can collect medication information about patients who have previously been prescribed ketamine as training data. Medication information may include the prescribed dosage for each patient, along with indicators of whether the patient's dosage was adjusted during treatment and / or whether the patient experienced adverse events. The drug and dosage determination server 102 can then analyze the training data to generate machine learning models (e.g., neural networks, decision trees, hyperplanes, regression models, etc.) to determine the new patient's dosage based on the new patient's biological, demographic, and clinical characteristics. Patient features utilized in the dosing algorithm may include biological data such as SNPs that have been reported to stratify human responses to ketamine. Patient features may also include demographic data such as the patient's sex, height and weight, age, and ethnicity. Furthermore, patient features may include clinical data such as family history, drug-drug interactions, history of mental illness, whether the patient smokes or uses nicotine, and Hamilton Depression Scale (HAM-D) scores.

[0109] In any case, the drug and dosage determination server 102 applies the patient's characteristics to the medication algorithm to determine the amount of ketamine to administer to the patient. The drug and dosage determination server 102 then provides the recommended dosage to the medical professional client devices 106-116.

[0110] While regression analyses of patient-specific dose optimization fail to account for nearly half of the pharmacogenetic and clinical variables necessary for accuracy, several published studies have reported variables to include in algorithmic decisions for antidepressant selection and dose estimation. As shown in Figure 13, clinical values ​​obtained from medical records are also important in determining the reduction of ketamine dosage. These include body mass index (BMI) above 30, family history of alcohol use disorder (grade 1), history of suicide attempts (multiple), previous psychiatric hospitalizations, female biological sex (premenopausal), non-Hispanic white ethnicity, and smoking habits. As shown, these values ​​may contribute to a substantial reduction in ketamine dosage, but do not rule out the use of this medication.

[0111] In addition to comparing the patient subnetwork for the target psychotropic drug with the reference subnetwork for the target psychotropic drug, the drug and dosage determination server 102 analyzes the patient's clinical data and neuroimaging data to determine whether or not to administer the drug to the patient. For example, the drug and dosage determination server 102 can analyze the patient's HAMD score and / or patient symptoms to classify the patient into one of four TRD patient subtypes.

[0112] Figure 19 shows four TRD patient subtypes determined by transcranial magnetic stimulation (TMS) and neuroimaging of the resting-state connectivity network in the human brain. In TRD patients, subtypes have been identified and replicated using TMS and neuroimaging studies, and the subtypes have been discretized by differential activation and inhibition of the resting-state connectivity network in the human brain. TMS has four different anatomical locations on the lateral side of the human head and activates various neuroanatomical structures that are part of the limbic cortical circuit, including the resting-state connectivity network, default mode network, defensive responses including the amygdala, reward circuits located in the basal forebrain including the nucleus accumbens (NA) or orbitofrontal cortex (OFC) in depression subtype 3, gating of sensory stimuli to the cortex via the thalamus, cortical region S1 and insula, and part of the prefrontal cortical inhibition of impulsivity, including the inhibition of amygdala activation in the limbic cortex, dorsolateral and dorsomedial prefrontal cortex, and memory enhancement in the entire anterior cingulate cortex and hippocampus.

[0113] Figure 19 also shows that TRD subtypes can be identified using structured values ​​or natural language processing from records, or using clinical data obtained from EHRs or other clinical records.

[0114] Figure 20 shows the brain regions that are consistently activated during the ketamine antidepressant response and how they map differently to four TRD depression subtypes. While brain regions are involved, these findings are consistent with both the different depression subtypes defined by TMS and clinical values, and the results of neuroimaging studies examining ketamine-induced activation and suppression.

[0115] Figure 21 shows recommendations for switching from ketamine for four different TRD subtypes, based on the HAMD-17 assessment (Hamilton Depression Scale), neuroimaging meta-analysis, relevant clinical values ​​shown in Figure 19, and available efficacy information, currently available antidepressant indications and recommendations, as well as expert guidelines from the American Psychiatric Association.

[0116] Figure 21 shows examples of recommended medications and alternative medication options for each of the four different subtypes of TRD depression patients.

[0117] To identify which depression subtypes should or should not be treated with ketamine and ketamine analogues, 24 publicly available neuroimaging datasets were analyzed to determine the neuroanatomical regions activated by ketamine and its analogues, as well as in patients with depression, TRD, and healthy controls. Patients with TRD subtype 3 consistently exhibit hyperactive subgenital anterior cingulate cortex (sgACC), dorsolateral and dorsomedral prefrontal (executive) cortex (dlPFC, dmPFC), and hyperactive orbitofrontal cortex (OFC), therefore, ketamine therapy is not recommended for these patients. This patient cohort neither responds nor achieves remission, but instead may experience exacerbation of psychotropic adverse drug events. Independent analysis of the neuroanatomical localization of all genes found in the ketamine pharmacogenetic network shows that they are all highly expressed in the anterior cingulate cortex, prefrontal cortex, supplemental motor cortex, orbitofrontal cortex, temporal cortex, amygdala, hippocampus, anterior caudate nucleus and nucleus accumbens (but not in other cortical brain regions), hypothalamus, or brainstem. This is consistent with the pattern observed in 24 functional neuroimaging studies examined, indicating where ketamine first exerts its antidepressant effects in the human brain (Table 1). [Table 1-1] [Table 1-2] [Table 1-3] [Table 1-4] [Table 1-5]

[0118] In another embodiment, disease risk and pharmacogenomic SNPs that distinguish two significantly different ketamine subnetworks in the human brain may be used to determine patient responses and adverse events when treated with ketamine. Table 2A shows enhancer and super-enhancer SNPs found in the ketamine efficacy subnetwork that can be used to determine the expression of mutations significantly associated with an effective response to ketamine. In contrast, Table 2B shows enhancer and super-enhancer SNPs found in the ketamine adverse events subnetwork that can be used to determine the expression of mutations significantly associated with adverse CNS events in response to ketamine. [Table 2-1] [Table 2-2] [Table 2-3] [Table 2-4] [Table 2-5] [Table 2-6] [Table 2-7] [Table 2-8] [Table 3-1] [Table 3-2] [Table 3-3] [Table 3-4] Table 3-5 Table 3-6 Table 3-7 Table 3-8 Table 3-9 Table 3-10 Table 3-11 Table 3-12 Table 3-13 Table 3-14 Table 3-15 Table 3-16 Table 3-17 Table 3-18 Table 3-19 Table 3-20 Table 3-21 Table 3-22 Table 3-23 Table 3-24 Table 3-25 Table 3-26 Table 3-27 Table 3-28 Table 3-29 Table 3-30 Table 3-31 Table 3-32 Table 4-1 Table 4-2 Table 4-3 Table 4-4 Table 5-1 Table 5-2 Table 5-3 Table 5-4 Table 5-5 Table 5-6 Table 5-7 Table 5-8 Table 6-1 Table 6-2 Table 6-3 Table 6-4 Table 6-5 Table 6-6 Table 6-7 Table 6-8 Table 6-9 Table 6-10 Table 6-11 Table 6-12 Table 6-13 Table 6-14 Table 6-15 Table 6-16 Table 6-17 Table 6-18 Table 6-19 Table 6-20 Table 6-21 Table 6-22 Table 6-23 Table 6-24 Table 6-25 Table 6-26 Table 6-27 Table 7-1 [Table 7-2] [Table 7-3] [Table 7-4] [Table 7-5]

[0119] In another embodiment of the method described herein, a generalization of this method can be used to identify FDA-approved drug combinations that can be used to enhance the treatment of neuropsychiatric disorders through corresponding network or subnetwork mechanisms in biology. Figure 22 illustrates an example of how the pharmacogenomic network of valproic acid and the pharmacogenomic network of ketamine function complementaryly to support neurogenesis. Valproic acid induces the conversion of neural progenitor cells to constrained neural progenitor cells via the npBAF complex (1, top), and ketamine may act on constrained neural progenitor cells via the human silencing complex (HUSH) to convert progenitor cells into differentiated neurons (1, bottom).

[0120] Figure 23 illustrates how the process shown in Figure 22 works through gradual deacetylation of the histone 3-lysine 9 (H3K9) moiety triggered by the valproic acid pharmacogenetic network (Figure 23A), and subsequent acetylation of the H3K9 moiety following activation of the ketamine pharmacogenetic network (Figure 23B).

[0121] Figure 24 illustrates how the complementary pharmacogenetic network of valproic acid and ketamine transforms neural progenitor cells into mature, differentiated neurons.

[0122] In another embodiment of this disclosure, these methods can be used for other antidepressants that target the NMDAR network. For example, other NMDAR partial antagonists such as AVP-786 and GLYX-13 (rapastinel) are being clinically tested as antidepressants. Also, GLRB blockers of the NMDAR are under development as antidepressants, such as AV101 and D-cycloserine (seromycin). Selective antagonists of the NMDAR GRIN2B are also under development as antidepressants, including EVT103, CP101, and MK-0657. Downstream of this pathway is AMPAR, and several antidepressants are being developed as GRIA1 and GRIA2 agonists, such as ORG265576.

[0123] In another embodiment, these methods can be used to optimize drug selection for other antidepressants, offering greater power than commercially available pharmacogenomic clinical decision support assays that rely solely on coated SNPs for classifying patients with respect to drugs. The methods encompassed in the techniques disclosed herein leverage pharmacogenomic epigenomic knowledge, including organization into TAD and TAD-TAD pharmacogenomic connections, to enhance insights into CNS drug mechanisms. Furthermore, these methods enable objective monitoring of drug-drug interactions and dosage, as well as measurement of parent drugs and their metabolites from serum, as in the case of S-ketamine and its active metabolite norketamine, providing additional insights into the subtypes of individual metabolizers.

[0124] Throughout this specification, multiple examples may implement components, operations, or structures described as a single example. While individual operations of one or more methods are illustrated and described as separate operations, one or more of these operations may be performed simultaneously, and they do not need to be performed in the illustrated order. Structures and functions presented as separate components within an illustrative configuration may be implemented as a combined structure or component. Similarly, structures and functions presented as single components may be implemented as separate components. These and other variations, modifications, additions, and improvements fall within the scope of the subject matter of this specification.

[0125] Furthermore, specific embodiments described herein include logic or a number of routines, subroutines, applications, or instructions. These can constitute either software (e.g., code embodied on a machine-readable medium or in a transmitted signal) or hardware. In hardware, routines, etc., are tangible units capable of performing specific operations and can be configured or arranged in specific ways. In exemplary embodiments, one or more computer systems (e.g., standalone, client, or server computer systems), or one or more hardware modules of a computer system (e.g., processors or groups of processors), can be configured by software (e.g., an application or part of an application) as hardware modules that operate to perform specific operations described herein.

[0126] In various embodiments, hardware modules can be implemented mechanically or electronically. For example, a hardware module may include a dedicated circuit or logic permanently configured to perform a specific operation (e.g., a field-programmable gate array (FPGA) or an application-specific integrated circuit (ASIC) or other application-specific processor). A hardware module may also include programmable logic or circuit temporarily configured by software to perform a specific operation (e.g., implemented in a general-purpose processor or other programmable processor). It will be understood that the decision of whether to implement a hardware module mechanically, with a dedicated and permanently configured circuit, or with a temporarily configured circuit (e.g., configured by software) can be made considering cost and time.

[0127] Therefore, the term “hardware module” should be understood to encompass tangible entities that are physically constructed, permanently configured (e.g., physically embedded), or temporarily configured (e.g., programmed) for the purpose of operating in a particular manner or performing a particular operation as described herein. When considering embodiments in which a hardware module is temporarily configured (e.g., programmed), each hardware module does not need to be configured or instantiated at any given time. For example, if a hardware module includes a general-purpose processor configured using software, that general-purpose processor can be configured as different hardware modules at different times. Thus, the software may configure the processor, for example, to configure one hardware module at one time and another hardware module at another time.

[0128] Hardware modules can provide information to other hardware modules and receive information from other hardware modules. Therefore, the hardware modules described herein can be understood as being communicatively coupled. When such hardware modules exist simultaneously, communication can be achieved through signal transmission (through appropriate circuits and buses) connecting the hardware modules. In embodiments where multiple hardware modules are configured or instantiated at different points in time, communication between such hardware modules can be achieved, for example, through the storage and retrieval of information in a memory structure accessible to the multiple hardware modules. For example, one hardware module may perform an operation and store the output of that operation in a memory device to which that hardware module is communicatively coupled. Further hardware modules can then access the memory device to retrieve and process the stored output. Hardware modules can also initiate communication with input or output devices to act on resources (e.g., information gathering).

[0129] Various operations of the exemplary methods described herein may be performed, at least in part, by one or more processors that are temporarily (e.g., by software) or permanently configured to perform the operations in question. Whether temporarily or permanently configured, such processors may constitute a processor implementation module that operates to perform one or more operations or functions. The modules referred to herein may include processor implementation modules in some exemplary embodiments.

[0130] Similarly, any method or routine described herein can be at least partially processor-implemented. For example, at least part of the operation of a method may be performed by one or more processors or processor-implemented hardware modules. Certain performance of the operation may reside not only within a single machine but also distributed among one or more processors deployed across several machines. In some embodiments, one or more processors may reside in a single location (e.g., in a home environment, a work environment, or as a server farm), while in other embodiments, the processors may be distributed across multiple locations.

[0131] Reliable performance can reside not only within a single machine but also distributed across one or more processors deployed across several machines. In some exemplary embodiments, one or more processors or processor implementation modules may reside in a single location (e.g., in a home environment, a work environment, or as a server farm). In other exemplary embodiments, one or more processors or processor implementation modules may be distributed across multiple locations.

[0132] Unless otherwise specified, any description in this specification using terms such as “process,” “calculate,” “calculate,” “determine,” “present,” or “display” may mean the operation or processing of a machine (e.g., a computer) that manipulates or transforms data expressed as physical (e.g., electronic, magnetic, or optical) quantities in one or more memories (e.g., volatile memory, non-volatile memory, or a combination thereof), registers, or other machine components that receive, store, transmit, or display information.

[0133] When used herein, any reference to “one embodiment” or “embodiment” means that the specific elements, features, structures, or characteristics described in conjunction with the embodiment are included in at least one embodiment. The phrase “in one embodiment” appearing in various places herein does not necessarily refer to the same embodiment.

[0134] Some embodiments may be described using the expressions “combined” and “connected,” along with their conjugations. For example, some embodiments may be described using the term “combined” to indicate that two or more elements are in direct physical or electrical contact. However, the term “combined” can also mean that two or more elements are not in direct contact with each other but are still cooperating or interacting with each other. Embodiments are not limited to this context.

[0135] When used herein, the terms “comprises,” “comprising,” “includes,” “including,” “has,” “having,” or any other variations thereof are intended to encompass non-exclusive inclusion. For example, a process, method, article, or apparatus that includes a list of elements is not necessarily limited to those elements alone, and may include other elements that are not expressly enumerated or that are inherent to such process, method, article, or apparatus. Furthermore, unless expressly stated to the contrary, “or” means an inclusive or rather an exclusive or. For example, condition A or B is satisfied by any one of the following: A is true (or exists) and B is false (or does not exist), A is false (or does not exist) and B is true (or exists), and both A and B are true (or exist).

[0136] In addition, the use of "a" or "an" is used to describe elements and components of embodiments herein. This is done solely for convenience and to provide a general summary of this specification. This description should be read as including one or at least one, and the singular includes the plural unless it is obvious that it is meant not to. The detailed description should be interpreted as merely providing examples, and does not describe all possible embodiments, as it would be impractical, if not impossible, to describe all possible embodiments. Numerous other embodiments may be carried out using either the current art or art developed after the filing date of this application.

Claims

1. A computing device for determining one or more drugs to be administered to a patient suffering from a neuropsychiatric disorder, wherein the computing device is Communication networks and One or more processors, The computing device comprises a non-temporary computer-readable memory coupled to one or more processors and storing instructions, and when executed by one or more processors according to the instructions, the computing device Data is obtained from the patient's biological samples, Obtain the reference drug pharmacological genomics network representation of glutamate N-methyl-d-aspartate receptor (NMDAR) antagonists or partial antagonists from a reference database. The data from the biological samples is analyzed considering the aforementioned reference drug pharmacological genomics network representation. Based on the above analysis, the pharmacological genomic network representation of patients treated with the NMDAR antagonist or partial antagonist was determined. A similarity score is determined by comparing the pharmacological genomics network representation of the patient with the pharmacological genomics network representation of the reference drug pharmacological genomics network representation of the NMDAR antagonist or partial antagonist. By analyzing the data from the biological sample using pharmacokinetic metabolomics analysis, the existing drug in the patient, the metabolite of the existing drug, and the dosage of the existing drug are determined. (i) based on the similarity score, and (ii) based on the drug-gene or drug-drug interaction between the patient's existing medications and the NMDAR antagonist or partial antagonist, a decision is made to administer the NMDAR antagonist or partial antagonist to the patient. The system displays an instruction to administer the NMDAR antagonist or partial antagonist, thereby ensuring that the NMDAR antagonist or partial antagonist is administered to the patient. A computing device in which the NMDAR antagonist or partial antagonist is administered to the patient.

2. The instruction further instructs the computing device to: From the aforementioned reference database, a constituent subnetwork is obtained for the NMDAR antagonist or partial antagonist, including at least the pharmacodynamic efficacy subnetwork and the pharmacodynamic adverse event subnetwork of the reference drug, wherein the first subset of variants in the pharmacodynamic adverse event subnetwork of the reference drug is causally related to the adverse events of the NMDAR antagonist or partial antagonist, and the second subset of variants in the pharmacodynamic efficacy subnetwork of the reference drug is causally related to the efficacy of the NMDAR antagonist or partial antagonist. The patient's pharmacological genomics network representation includes a pharmacodynamic efficacy subnetwork for the patient's drugs and a pharmacodynamic adverse events subnetwork for the patient's drugs, and in order to determine the similarity score, the instruction is given to the computing device, Comparison of the pharmacodynamic efficacy subnetwork of the reference drug and the pharmacodynamic efficacy subnetwork of the patient's drug, or Comparison of the pharmacodynamic adverse event subnetwork of the reference drug and the pharmacodynamic adverse event subnetwork of the patient's drug. The computing device according to claim 1, wherein the similarity score is determined according to at least one of the following.

3. The pharmacological genomics network representation of the patient for the NMDAR antagonist or partial antagonist includes a pharmacodynamic efficacy subnetwork for the patient, a pharmacodynamic adverse events subnetwork for the patient, a chromatin remodeling subnetwork, and a reference pharmacokinetic enzyme and hormone subnetwork; the constituent subnetwork of the NMDAR antagonist or partial antagonist from the reference database includes a pharmacodynamic efficacy subnetwork for the reference drug, a pharmacodynamic adverse events subnetwork for the reference drug, a reference chromatin remodeling subnetwork, and a reference pharmacokinetic enzyme and hormone subnetwork spanning human drug response variations; In order to determine the similarity score, the instruction causes the computing device to: Based on the amount of similarity between the pharmacodynamic efficacy subnetwork of the patient's drug and the pharmacodynamic efficacy subnetwork of the reference drug, a first score is assigned to the pharmacodynamic efficacy subnetwork of the patient's drug. The computing device according to claim 2, which assigns a second score to the pharmacodynamic adverse event subnetwork of the patient's drug based on the amount of similarity between the subnetwork of the patient's drug and the pharmacodynamic adverse event subnetwork of the reference drug.

4. In order to decide to administer the NMDAR antagonist or partial antagonist to the patient based on the similarity score, the computing device, by order of the instruction, The computing device according to claim 3, which determines to administer the NMDAR antagonist or partial antagonist to the patient if the first score exceeds a first threshold score or the second score falls below a second threshold score.

5. When the first score exceeds a first threshold score or the second score falls below a second threshold score, the computing device, in order to decide to administer the NMDAR antagonist or partial antagonist to the patient, the instruction causes the computing device to: The similarity score is determined by combining the first and second scores. The computing device according to claim 4, wherein if the similarity score exceeds a third threshold score, it is determined to administer the NMDAR antagonist or partial antagonist to the patient.

6. The aforementioned instruction further causes the computing device to: Based on the similarity score, the dosage of the NMDAR antagonist or partial antagonist to be administered to the patient is determined. The computing device according to claim 1, which causes the patient to administer the determined dose of the NMDAR antagonist or partial antagonist by displaying an instruction to administer the determined dose of the NMDAR antagonist or partial antagonist.

7. In order to determine the dosage of the NMDAR antagonist or partial antagonist to be administered to the patient, the computing device, by order of the instruction, The computing device according to claim 6, which determines the dosage using a regression model based on two or more combinations of the patient's sex, the patient's age, whether or not the patient smokes, the patient's ethnicity, the patient's height, the patient's weight, and the patient's history of mental illness.

8. The computing device according to claim 1, wherein the reference drug pharmacological genomics network representation of the NMDAR antagonist or partial antagonist obtained from the reference database is a ketamine pharmacological genomics network representation, and comprises one or more of the following: a regulatory cytoskeleton-related protein (ARC) gene, an Achaete-Scute family bHLH transcription factor 1 (ASCL1) gene, a brain-derived neurotrophic factor (BDNF) gene, and a BDNF antisense RNA (BDNF-AS) gene. , Calcium / calmodulin-dependent protein kinase II alpha (CAMK2A) gene, Cyclin-dependent kinase inhibitor 1A (CDKN1A) gene, cAMP response element modulator (CREM) gene, Cut-like homeobox 2 (CUX2) gene, DCC netrin 1 receptor (DCC) gene, Dopamine receptor D2 (DRD2) gene, Fragile X intellectual disability 1 (FMR1) gene, Ganglioside-induced differentiation-related protein 1-like 1 (GDPP1L1) gene, Glutamate metabolism receptor 5 (GRM5) gene The genes include: Homer scaffold protein 1 (HOMER1) gene, 5-hydroxytryptamine receptor 1B (HTR1B) gene, 5-hydroxytryptamine receptor 2A (HTR2A) gene, Kruppel-like factor 6 (KLF6) gene, Lin-7 homolog C, long non-coding RNA of crumbs cell polarity complex component (LIN7C), long non-coding RNA of LOC105379109, myocyte enhancer factor 2D (MEF2D) gene, myosin VI (MYO6) gene, myelin transcription factor 1-like (MYT1L) ) gene, neural differentiation 1 (NEUROD1) gene, neural differentiation 2 (NEUROD2) gene, neonthic helix-loop-helix 2 (NHLH2) gene, neuromedin B (NMB) gene, NMDA receptor synaptic nuclear signaling and neuronal migration factor (NSMF) gene, neurotrophic receptor tyrosine kinase 2 (NTRK2) gene, phosphotase and tensin homolog (PTEN) gene, prostaglandin-endoperoxide synthase 2 (PTGS2) gene, Rac family small molecule GTPase 1 (RAC1) gene, Ras protein-specific guanine nucleotide release factor 2 (RASGRF2) gene,Ras homolog family member A (RHOA) gene, roundabout guidance receptor 2 (ROBO2) gene, RP11_360A181 long non-coding RNA, semaphorin 3A (SEMA3A) gene, SH3 and multiple ankyrin repeat domain 1 (SHANK1) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) gene, SH3 and multiple ankyrin repeat domain 3 (SHANK3) gene, solute carrier family 22 member 15 (SLC22A15) gene, solute carrier family 6 member 2 (SLC6A 2) Genes, Slit Guidance Ligand 1 (SLIT1) gene, Slit Guidance Ligand 2 (SLIT2) gene, Synaptosome-related protein 25 (SNAP25) gene, Synapsin I (SYN1) gene, Synapsin II (SYN2) gene, Synapsin III (SYN3) gene, T-box, Brain 1 (TBR1) gene, Transcription factor 4 (TCF4) gene, Acetylcholinesterase (ACHE) gene, Activating Transcription Factor 7 Interacting Protein (ATF7IP) gene, Activating Transcription Factor 7 Interacting Protein 2 (ATF7IP2) gene, ATPase The genes for the Na+ / K+ transport subunit alpha-1 (ATP1A1), BLOC-1-related complex unit 7 (BORCS7), bromodomain-containing 4 (BRD4), calcium voltage-gated channel subunit alpha-1C (CACNA1C), calcium voltage-gated channel accessory subunit beta-1 (CACNB1), calcium voltage-gated channel accessory subunit beta-2 (CACNB2), and calcium voltage-gated channel accessory subunit gamma-2 (CACNG2) are also included. The genes for the cholinergic receptor muscarinic 2 (CHRM2), cholinergic receptor nicotinic alpha 3 subunit (CHRNA3), cholinergic receptor nicotinic alpha 5 subunit (CHRNA5), cholinergic receptor nicotinic alpha 7 subunit (CHRNA7), cannabinoid receptor 1 (CNR1), discrepant homolog 3 (DLG3), discrepant homolog 4 (DLG4), DNA methyltransferase 1 (DNMT1),Euchromatic histone lysine methyltransferase 1 (EHMT1) gene, gamma-aminobutyric acid type A receptor alpha 2 subunit (GABRA2) gene, gamma-aminobutyric acid type A receptor alpha 5 subunit (GABRA5) gene, glutamate decarboxylase 1 (GAD1) ​​gene, glycine receptor alpha 1 (GLRA1) gene, glycine receptor alpha 2 (GLRA2) gene, glycine receptor beta (GLRB) gene, glutamate ion channel embedded receptor AMPA subunit 1 (GRIA1) Genes, Glutamate ion channel-embedded receptor AMPA subunit 2 (GRIA2) gene, Glutamate ion channel-embedded receptor AMPA subunit 4 (GRIA4) gene, Glutamate ion channel-embedded receptor NMDA subunit 1 (GRIN1) gene, Glutamate ion channel-embedded receptor NMDA subunit 2A (GRIN2A) gene, Glutamate ion channel-embedded receptor NMDA subunit 2B (GRIN2B) gene, Glutamate ion channel-embedded receptor NMDA subunit GRIN2C gene, Glutamate ion channel embedded receptor NMDA subunit 2D gene, GRIN3A gene, Glutamate ion channel embedded receptor NMDA subunit 3B gene, Hyperpolarization-activated cyclic nucleotide-gate potassium channel 1 (HCN1) gene, Histone deacetylase 5 (HDAC5) gene, Methyl-CpG-binding domain protein 1 (MBD1) gene, M-phase phosphorylated protein 8 (MPHOSPH8) gene, neuronal cell adhesion molecule 1 (NCAM1) gene, nitrate synthase 1 (NOS1) gene, nitrate synthase 2 (NOS2) gene, nitrate synthase 3 (NOS3) gene, NAD(P)H quinone dehydrogenase 1 (NQO1) gene, opioid receptor kappa 1 (OPRK1) gene, opioid receptor mu 1 (OPRM1) gene, roundabout guidance receptor 2 (ROBO2) gene, SET domain branch 1 (SETDB1) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) genes,The genes include the Sigma non-opioid intracellular receptor 1 (SIGMAR1), solute carrier family 6 member 9 (SLC6A9), transcriptional activation suppressor (TASOR) gene, and the TOG array regulator of axonal microtubules. 2) The (TOGORAM2) gene, the TRIM28 gene containing three-element motifs, the ZNF274 gene, the ANAPC2 gene, the cytochrome P450 family 2 subfamily A member 6 (CYP2A6) gene, the cytochrome P450 family 2 subfamily B member 6 (CYP2B6) gene, the cytochrome P450 family 3 subfamily A member 4 (CYP3A4) gene, the eukaryotic elongation factor 2 kinase (EEF2K) gene, the estrogen receptor 1 (ESR1) gene, or the transcription elongation regulator 1 (TCERG1) gene.

9. A method for determining drug administration for patients suffering from neuropsychiatric disorders, Obtaining biological samples from patients, Obtain the reference drug pharmacological genomics network representation of a glutamate N-methyl-d-aspartate receptor (NMDAR) antagonist or partial antagonist from a reference database using one or more processors. Analyzing the biological sample while considering the pharmacological genomics network representation of the reference drug, Based on the above analysis, the pharmacological genomics network representation of the patient with the NMDAR antagonist or partial antagonist is determined by one or more processors. The similarity score is determined by one or more processors based on a comparison between the pharmacological genomics network representation of the patient and the reference drug pharmacological genomics network representation of the NMDAR antagonist or partial antagonist. The clinical data of the aforementioned patient is acquired by one or more processors. By analyzing data from the biological sample using pharmacokinetic metabolomics analysis with one or more processors, the existing drug in the patient, the metabolites of the existing drug, and the dosage of the existing drug are determined. A method comprising: (i) determining, by one or more processors, whether to administer the NMDAR antagonist or partial antagonist to the patient based on the similarity score and (ii) the drug-gene or drug-drug interaction between the patient's existing medications and the NMDAR antagonist or partial antagonist.

10. The process further includes obtaining, from the aforementioned reference database, a constituent subnetwork for the NMDAR antagonist or partial antagonist, which includes at least the pharmacodynamic efficacy subnetwork and the pharmacodynamic adverse events subnetwork of the reference drug, using one or more processors, The first subset of variants in the pharmacodynamic adverse event subnetwork of the reference drug is causally related to the adverse events of the NMDAR antagonist or partial antagonist, and the second subset of variants in the pharmacodynamic efficacy subnetwork of the reference drug is causally related to the efficacy of the NMDAR antagonist or partial antagonist. The aforementioned patient pharmacological genomics network representation includes a subsystem for the patient's drug pharmacodynamic efficacy and a subsystem for the patient's drug pharmacodynamic adverse events. In determining the aforementioned similarity score, Comparison of the pharmacodynamic efficacy subnetwork of the reference drug and the pharmacodynamic efficacy subnetwork of the patient's drug, or Comparison of the pharmacodynamic adverse event subnetwork of the reference drug and the pharmacodynamic adverse event subnetwork of the patient's drug. The process includes determining the similarity score by one or more processors according to at least one of the following: The method according to claim 9.

11. The patient pharmacological genomics network representation of the NMDAR antagonist or partial antagonist includes a pharmacodynamic efficacy subsystem of the patient, a pharmacodynamic adverse events subsystem of the patient, a chromatin remodeling subsystem, and a pharmacokinetic enzyme and hormone subsystem; the constituent subsystem of the NMDAR antagonist or partial antagonist from the reference database includes a pharmacodynamic efficacy subsystem of the reference drug, a pharmacodynamic adverse events subsystem of the reference drug, a reference chromatin remodeling subsystem, and a reference pharmacokinetic enzyme and hormone subsystem spanning human drug response variations; In determining the similarity score, Based on the amount of similarity between the pharmacodynamic efficacy subnetwork of the patient's drug and the pharmacodynamic efficacy subnetwork of the reference drug, one or more processors assign a first score to the pharmacodynamic efficacy subnetwork of the patient's drug. The method according to claim 10, comprising assigning a second score to the pharmacodynamic adverse event subnetwork of the patient's drug by one or more processors, based on the amount of similarity between the pharmacodynamic adverse event subnetwork of the patient's drug and the pharmacodynamic adverse event subnetwork of the reference drug.

12. Based on the similarity score, the decision to administer the NMDAR antagonist or partial antagonist to the patient is made as follows: The method according to claim 11, comprising determining by one or more processors whether to administer the NMDAR antagonist or partial antagonist to the patient if the first score exceeds a first threshold score or the second score falls below a second threshold score.

13. If the first score exceeds the first threshold score, or if the second score falls below the second threshold score, the decision to administer the NMDAR antagonist or partial antagonist to the patient is made as follows: The first and second scores are combined by one or more processors to determine the similarity score, The method according to claim 12, comprising determining by one or more processors whether to administer the NMDAR antagonist or partial antagonist to the patient if the similarity score exceeds a third threshold score.

14. The method according to claim 9, further comprising determining, by the one or more processors, the dose of the NMDAR antagonist or partial antagonist to be administered to the patient based on the similarity score.

15. The determination of the dosage of the NMDAR antagonist or partial antagonist to be administered to the patient by one or more processors is as follows: The method according to claim 14, comprising determining the dosage using a regression model based on two or more combinations of the patient's sex, the patient's age, whether the patient smokes or not, the patient's ethnicity, the patient's height, the patient's weight, and the patient's history of mental illness.

16. The method according to claim 10, wherein the reference drug pharmacological genomics network representation of the NMDAR antagonist or partial antagonist obtained from the reference database is a ketamine pharmacological genomics network representation, and comprises one or more of the following: a regulatory cytoskeleton-related protein (ARC) gene, an Achaete-Scute family bHLH transcription factor 1 (ASCL1) gene, a brain-derived neurotrophic factor (BDNF) gene, a BDNF antisense RNA (BDNF-AS) gene, and a calcium / calcium / calcium-based cytoskeleton-related protein (ARC) gene. Lumodulin-dependent protein kinase II alpha (CAMK2A) gene, cyclin-dependent kinase inhibitor 1A (CDKN1A) gene, cAMP response element modulator (CREM) gene, Cut-like homeobox 2 (CUX2) gene, DCC netrin 1 receptor (DCC) gene, dopamine receptor D2 (DRD2) gene, fragile X intellectual disability 1 (FMR1) gene, ganglioside-induced differentiation-related protein 1-like 1 (GDPP1L1) gene, glutamate receptor 5 (GRM5) gene, homeobox 2 (CUX2) gene, - Scaffold protein 1 (HOMER1) gene, 5-hydroxytryptamine receptor 1B (HTR1B) gene, 5-hydroxytryptamine receptor 2A (HTR2A) gene, Kruppel-like factor 6 (KLF6) gene, Lin-7 homolog C, long non-coding RNA of crumbs cell polarity complex component (LIN7C), long non-coding RNA of LOC105379109, muscle cell enhancer factor 2D (MEF2D) gene, myosin VI (MYO6) gene, myelin transcription factor 1-like (MYT1L) gene The genes include: Neuron differentiation 1 (NEUROD1) gene, Neuron differentiation 2 (NEUROD2) gene, Nesient helix-loop-helix 2 (NHLH2) gene, Neuromedin B (NMB) gene, NMDA receptor synaptic nuclear signaling and neuronal migration factor (NSMF) gene, Neurotrophic receptor tyrosine kinase 2 (NTRK2) gene, Phosphotase and tensin homolog (PTEN) gene, Prostaglandin-endoperoxide synthase 2 (PTGS2) gene, Rac family small molecule GTPase 1 (RAC1) gene, Ras protein-specific guanine nucleotide release factor 2 (RASGRF2) gene,Ras homolog family member A (RHOA) gene, roundabout guidance receptor 2 (ROBO2) gene, RP11_360A181 long non-coding RNA, semaphorin 3A (SEMA3A) gene, SH3 and multiple ankyrin repeat domain 1 (SHANK1) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) gene, SH3 and multiple ankyrin repeat domain 3 (SHANK3) gene, solute carrier family 22 member 15 (SLC22A15) gene, solute carrier family 6 member 2 (SLC6A 2) Genes, Slit Guidance Ligand 1 (SLIT1) gene, Slit Guidance Ligand 2 (SLIT2) gene, Synaptosome-related protein 25 (SNAP25) gene, Synapsin I (SYN1) gene, Synapsin II (SYN2) gene, Synapsin III (SYN3) gene, T-box, Brain 1 (TBR1) gene, Transcription factor 4 (TCF4) gene, Acetylcholinesterase (ACHE) gene, Activating Transcription Factor 7 Interacting Protein (ATF7IP) gene, Activating Transcription Factor 7 Interacting Protein 2 (ATF7IP2) gene, ATPase The genes for the Na+ / K+ transport subunit alpha-1 (ATP1A1), BLOC-1-related complex unit 7 (BORCS7), bromodomain-containing 4 (BRD4), calcium voltage-gated channel subunit alpha-1C (CACNA1C), calcium voltage-gated channel accessory subunit beta-1 (CACNB1), calcium voltage-gated channel accessory subunit beta-2 (CACNB2), and calcium voltage-gated channel accessory subunit gamma-2 (CACNG2) are also included. The genes for the cholinergic receptor muscarinic 2 (CHRM2), cholinergic receptor nicotinic alpha 3 subunit (CHRNA3), cholinergic receptor nicotinic alpha 5 subunit (CHRNA5), cholinergic receptor nicotinic alpha 7 subunit (CHRNA7), cannabinoid receptor 1 (CNR1), discrepant homolog 3 (DLG3), discrepant homolog 4 (DLG4), DNA methyltransferase 1 (DNMT1),Euchromatic histone lysine methyltransferase 1 (EHMT1) gene, gamma-aminobutyric acid type A receptor alpha 2 subunit (GABRA2) gene, gamma-aminobutyric acid type A receptor alpha 5 subunit (GABRA5) gene, glutamate decarboxylase 1 (GAD1) ​​gene, glycine receptor alpha 1 (GLRA1) gene, glycine receptor alpha 2 (GLRA2) gene, glycine receptor beta (GLRB) gene, glutamate ion channel embedded receptor AMPA subunit 1 (GRIA1) Genes, Glutamate ion channel-embedded receptor AMPA subunit 2 (GRIA2) gene, Glutamate ion channel-embedded receptor AMPA subunit 4 (GRIA4) gene, Glutamate ion channel-embedded receptor NMDA subunit 1 (GRIN1) gene, Glutamate ion channel-embedded receptor NMDA subunit 2A (GRIN2A) gene, Glutamate ion channel-embedded receptor NMDA subunit 2B (GRIN2B) gene, Glutamate ion channel-embedded receptor NMDA subunit GRIN2C gene, Glutamate ion channel embedded receptor NMDA subunit 2D gene, GRIN3A gene, Glutamate ion channel embedded receptor NMDA subunit 3B gene, Hyperpolarization-activated cyclic nucleotide-gate potassium channel 1 (HCN1) gene, Histone deacetylase 5 (HDAC5) gene, Methyl-CpG-binding domain protein 1 (MBD1) gene, M-phase phosphorylated protein 8 (MPHOSPH8) gene, neuronal cell adhesion molecule 1 (NCAM1) gene, nitrate synthase 1 (NOS1) gene, nitrate synthase 2 (NOS2) gene, nitrate synthase 3 (NOS3) gene, NAD(P)H quinone dehydrogenase 1 (NQO1) gene, opioid receptor kappa 1 (OPRK1) gene, opioid receptor mu 1 (OPRM1) gene, roundabout guidance receptor 2 (ROBO2) gene, SET domain branch 1 (SETDB1) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) genes,The genes include the Sigma non-opioid intracellular receptor 1 (SIGMAR1), solute carrier family 6 member 9 (SLC6A9), transcriptional activation suppressor (TASOR) gene, and the TOG array regulator of axonal microtubules. 2) The (TOGORAM2) gene, the TRIM28 gene containing three-element motifs, the ZNF274 gene, the ANAPC2 gene, the cytochrome P450 family 2 subfamily A member 6 (CYP2A6) gene, the cytochrome P450 family 2 subfamily B member 6 (CYP2B6) gene, the cytochrome P450 family 3 subfamily A member 4 (CYP3A4) gene, the eukaryotic elongation factor 2 kinase (EEF2K) gene, the estrogen receptor 1 (ESR1) gene, or the transcription elongation regulator 1 (TCERG1) gene.

17. The method according to claim 16, wherein the pharmacodynamic efficacy subnetwork of the reference drug comprises one or more of the following: a regulatory cytoskeleton-related protein (ARC) gene, an Achaete-Scute family bHLH transcription factor 1 (ASCL1) gene, a brain-derived neurotrophic factor (BDNF) gene, a BDNF antisense RNA (BDNF-AS) gene, a calcium / calmodulin-dependent protein kinase II alpha (CAMK2A) gene, and a cyclin-dependent kinase inhibitor 1A (CDK N1A) gene, cAMP response element modulator (CREM) gene, Cut-like homeobox 2 (CUX2) gene, DCC netrin 1 receptor (DCC) gene, dopamine receptor D2 (DRD2) gene, eukaryotic elongation factor 2 kinase (EEF2K) gene, fragile X intellectual disability 1 (FMR1) gene, ganglioside-induced differentiation-related protein 1-like 1 (GDPR1L1) gene, glutamate receptor 5 (GRM5) gene, Homer scaffold protein 1 (HOMER1) gene, 5-hydroxytryptamine receptor 1B (HTR1B) gene, 5-hydroxytryptamine receptor 2A (HTR2A) gene, Kruppel-like factor 6 (KLF6) gene, Lin-7 homolog C, long non-coding RNA of crumbs cell polarity complex component (LIN7C), long non-coding RNA of LOC105379109, myocyte enhancer factor 2D (MEF2D) gene, myosin VI (MYO6) gene, myelin transcription factor 1-like (MYT1L) gene, neural differentiation 1 (NE UROD1 gene, Neuron Differentiation 2 (NEUROD2) gene, Nesient Helix Loop Helix 2 (NHLH2) gene, Neuromedin B (NMB) gene, NMDA receptor synaptic nuclear signaling and neuronal migration factor (NSMF) gene, Neurotrophic receptor tyrosine kinase 2 (NTRK2) gene, Phosphotase and Tensin Homolog (PTEN) gene, Prostaglandin-Endoperoxide Synthase 2 (PTGS2) gene, Rac family small molecule GTPase1 (RAC1) gene, Ras protein-specific guanine nucleotide release factor 2 (RASGRF2) gene, Ras homolog family member A (RHOA) gene, roundabout guidance receptor 2 (ROBO2) gene, long non-coding RNA of RP11_360A181, semaphorin 3A (SEMA3A) gene, SH3 and multiple ankyrin repeat domain 1 (SHANK1) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) gene, SH3 and multiple ankyrin repeat domain 3 (SHANK3) gene, solute carrier family 22 member 15 (SLC22A15) gene, solute carrier family 6 member 2 (SLC6A2) gene, slit guidance ligand 1 (SLIT1) gene, slit guidance ligand 2 (SLIT2) gene, synaptosome-related protein 25 (SNAP25) gene, synapsin I (SYN1) gene, synapsin II (SYN2) gene, synapsin III (SYN3) gene, T-box, brain 1 (TBR1) gene, or transcription factor 4 (TCF4) gene.

18. The method according to claim 16, wherein the pharmacodynamic adverse event subnetwork of the reference drug comprises one or more of the following: acetylcholinesterase (ACHE) gene, activator transcription factor 7 interacting protein (ATF7IP) gene, activator transcription factor 7 interacting protein 2 (ATF7IP2) gene, ATPase Na+ / K+ transport subunit alpha-1 (ATP1A1) gene, BLOC-1 related complex unit 7 (BORCS7) gene, bromodomain-containing 4 (BRD4) gene, calcium voltage-gated channel subunit alpha-1C (CACNA1C) gene, calcium voltage-gated channel co-subunit beta-1 (CACNB1) gene, calcium voltage-gated channel co-subunit beta-2 (CACNB2) gene, calcium voltage-gated channel co-subunit gamma-2 (CACNG2) gene, cholinergic receptor muscarinic 2 (CHRM2) gene, cholinergic receptor nicotinic alpha-3 subunit (CHRNA3) gene, cholinergic receptor nicotinic alpha-5 subunit (CHRNA5) gene, cholinergic receptor nicotinic alpha-7 subunit (CHRNA7) gene, cannabinoid receptor 1 (CNR1) gene, discreet homolog 3 (DLG3) gene, discreet Homolog 4 (DLG4) gene, DNA methyltransferase 1 (DNMT1) gene, euchromatic histone lysine methyltransferase 1 (EHMT1) gene, gamma-aminobutyric acid type A receptor alpha 2 subunit (GABRA2) gene, gamma-aminobutyric acid type A receptor alpha 5 subunit (GABRA5) gene, glutamate decarboxylase 1 (GAD1) ​​gene, glycine receptor alpha 1 (GLRA1) gene, glycine The genes for the glycine receptor alpha-2 (GLRA2), glycine receptor beta (GLRB), glutamate ion channel-embedded receptor AMPA subunit 1 (GRIA1), glutamate ion channel-embedded receptor AMPA subunit 2 (GRIA2), glutamate ion channel-embedded receptor AMPA subunit 4 (GRIA4), and glutamate ion channel-embedded receptor NMDA subunit 1 (GRIN1),Glutamate ion channel embedded receptor NMDA subunit 2A (GRIN2A) gene, glutamate ion channel embedded receptor NMDA subunit 2B (GRIN2B) gene, glutamate ion channel embedded receptor NMDA subunit 2C (GRIN2C) gene, glutamate ion channel embedded receptor NMDA subunit 2D (GRIN2D) gene, glutamate ion channel embedded receptor NMDA subunit 3A (GRIN3A) gene, glutamate ion channel embedded receptor NMDA subunit 3B (GRIN3B) gene, hyperpolarization-activated cyclic nucleotide-gate potassium channel 1 (HCN1) gene, histone deacetylase 5 (HDAC5) gene, methyl-CpG-binding domain protein 1 (MBD1) gene, M-phase phosphate The genes include the MPHOSPH8 protein 8 gene, the NCAM1 neuronal adhesion molecule 1 gene, the NOS1 nitrate synthase 1 gene, the NOS2 nitrate synthase 2 gene, the NOS3 nitrate synthase 3 gene, the NAD(P)H quinone dehydrogenase 1 gene (NQO1 gene), the opioid receptor kappa 1 gene (OPRK1 gene), the opioid receptor mu 1 gene (OPRM1 gene), the roundabout guidance receptor 2 gene (ROBO2 gene), the SET domain branch 1 gene (SETDB1 gene), the SH3 and multiple ankyrin repeat domain 2 (SHANK2) gene, the Sigma nonopioid intracellular receptor 1 (SIGMAR1) gene, the solute carrier family 6 member 9 (SLC6A9) gene, the transcriptional activation suppressor (TASOR) gene, and the TOG array regulator 2 (TOG) for axonal microtubules. The array regulator of axonemal microtubules 2) (TOGORAM2) gene, the tri-element motif-containing 28 (TRIM28) gene, or the zinc finger protein 274 (ZNF274) gene.

19. The method according to claim 16, wherein the constituent subnetwork of the ketamine pharmacological genomics network representation includes a reference pharmacokinetic enzyme and hormone subnetwork, and the reference pharmacokinetic enzyme and hormone subnetwork includes one or more of the following: late-stage acceleration complex subunit 2 (ANAPC2) gene, cytochrome P450 family 2 subfamily A member 6 (CYP2A6) gene, cytochrome P450 family 2 subfamily B member 6 (CYP2B6) gene, cytochrome P450 family 3 subfamily A member 4 (CYP3A4) gene, disc large homolog 4 (DLG4), eukaryotic elongation factor 2 kinase (EEF) The 2K gene, estrogen receptor 1 (ESR1) gene, glutamate ion channel-embedded receptor AMPA subunit 1 (GRIA1) gene, glutamate ion channel-embedded receptor AMPA subunit 4 (GRIA4) gene, glutamate ion channel-embedded receptor NMDA subunit 1 (GRIN1) gene, glutamate ion channel-embedded receptor NMDA subunit 2B (GRIN2B) gene, myosin VI (MYO6) gene, roundabout guidance receptor 2 (ROBO2) gene, SH3 and multiple ankyrin repeat domain 2 (SHANK2) genes, or transcription elongation regulator 1 (TCERG1) gene.

Images