Advertisement

Lessons learned from parsing genetic risk for schizophrenia into biological pathways

  • Author Footnotes
    ∗ denotes equal contribution
    Giulio Pergola
    Correspondence
    Corresponding author: Giulio Pergola, PhD, Professor of Biological Psychology, Department of Basic Medical Sciences, Neuroscience, and Sense Organs, University of Bari Aldo Moro, I-70121 Bari, Italy, Tel.: +39 0805478548. Emails to:
    Footnotes
    ∗ denotes equal contribution
    Affiliations
    Department of Basic Medical Sciences, Neuroscience, and Sense Organs, University of Bari Aldo Moro
    Search for articles by this author
  • Author Footnotes
    ∗ denotes equal contribution
    Nora Penzel
    Footnotes
    ∗ denotes equal contribution
    Affiliations
    Department of Basic Medical Sciences, Neuroscience, and Sense Organs, University of Bari Aldo Moro
    Search for articles by this author
  • Leonardo Sportelli
    Affiliations
    Department of Basic Medical Sciences, Neuroscience, and Sense Organs, University of Bari Aldo Moro
    Search for articles by this author
  • Alessandro Bertolino
    Affiliations
    Department of Basic Medical Sciences, Neuroscience, and Sense Organs, University of Bari Aldo Moro
    Search for articles by this author
  • Author Footnotes
    ∗ denotes equal contribution
Open AccessPublished:October 27, 2022DOI:https://doi.org/10.1016/j.biopsych.2022.10.009

      Abstract

      The clinically heterogeneous presentation of schizophrenia is compounded by the heterogeneity of risk factors and neurobiological correlates of the disorder. Genome-wide association studies in schizophrenia have uncovered a remarkably high number of genetic variants, but the biological pathways they impact upon remain largely unidentified. Among the diverse methodological approaches employed to provide a more granular understanding of genetic risk for schizophrenia, the use of biological labels such as gene ontologies, regulome approaches, and gene co-expression, have all provided novel perspectives into how genetic risk translates into the neurobiology of schizophrenia. Here, we review the salient aspects of parsing polygenic risk for schizophrenia into biological pathways. We argue that parsed scores, compared to standard polygenic risk scores, may afford more biologically plausible and accurate physiological modeling of the different dimensions involved in translating genetic risk into brain mechanisms, including multiple brain regions, cell types, and maturation stages. We discuss caveats, opportunities, and pitfalls inherent in the parsed risk approach.

      Keywords

      The biology of major psychiatric disorders is rooted in genetic risk (
      • Gottesman II,
      • Shields J.
      A Polygenic Theory of Schizophrenia.
      ,
      • Gottesman II,
      • Gould T.D.
      The Endophenotype Concept in Psychiatry: Etymology and Strategic Intentions.
      ,
      • Sullivan P.F.
      • Kendler K.S.
      • Neale M.C.
      Schizophrenia as a Complex Trait:. Evidence From a Meta-analysis of Twin Studies.
      ,
      • Kendler K.S.
      What psychiatric genetics has taught us about the nature of psychiatric illness and what is left to learn.
      ). Identifying neuropathological signatures shared by patients with the same psychiatric diagnosis has proven more challenging than for most brain disorders with a genetic component (
      • Weinberger D.R.
      • Radulescu E.
      Finding the Elusive Psychiatric "Lesion" With 21st-Century Neuroanatomy: A Note of Caution.
      ). Some of the most convincing results have been found in schizophrenia (SCZ; Figure 1) (
      • Sullivan P.F.
      • Geschwind D.H.
      Defining the Genetic, Genomic, Cellular, and Diagnostic Architectures of Psychiatric Disorders.
      ). Likely because of the high heritability estimates of 60-80 % (
      • Weinberger D.R.
      Thinking About Schizophrenia in an Era of Genomic Medicine.
      ,

      Sullivan PF, Kendler KS, Neale MC: Schizophrenia as a Complex Trait: Evidence From a Meta-analysis of Twin Studies.

      ,

      Uher R, Zwicker A: Etiology in psychiatry: embracing the reality of poly‐gene‐environmental causation of mental illness.

      ) and the high number of cases and controls available in the latest genome-wide association study (GWAS) (
      • Trubetskoy V.
      • Pardiñas A.F.
      • Qi T.
      • Panagiotaropoulou G.
      • Awasthi S.
      • Bigdeli T.B.
      • et al.
      Mapping genomic loci implicates genes and synaptic biology in schizophrenia.
      ), the last decade has delivered remarkable progress in understanding SCZ genetics. The latest GWAS of the Psychiatric Genetics Consortium (PGC; (
      • Trubetskoy V.
      • Pardiñas A.F.
      • Qi T.
      • Panagiotaropoulou G.
      • Awasthi S.
      • Bigdeli T.B.
      • et al.
      Mapping genomic loci implicates genes and synaptic biology in schizophrenia.
      )) has reported common variant associations with SCZ in 287 distinct loci. Each variant only explains a tiny proportion of risk for SCZ, but when combined into composite scores may yield odds ratios of 39 in the comparison between top and bottom centile (
      • Trubetskoy V.
      • Pardiñas A.F.
      • Qi T.
      • Panagiotaropoulou G.
      • Awasthi S.
      • Bigdeli T.B.
      • et al.
      Mapping genomic loci implicates genes and synaptic biology in schizophrenia.
      ). However, a large gap between the estimation of twin- versus genotype-based heritability remains (
      • Sullivan P.F.
      • Geschwind D.H.
      Defining the Genetic, Genomic, Cellular, and Diagnostic Architectures of Psychiatric Disorders.
      ,

      Uher R, Zwicker A: Etiology in psychiatry: embracing the reality of poly‐gene‐environmental causation of mental illness.

      ). This heritability gap suggests a complex genetic architecture, a potential role of non-additive genetics and gene-environment interplay, heterogeneous etiologies and clinical manifestations, or a combination of these factors (

      Uher R, Zwicker A: Etiology in psychiatry: embracing the reality of poly‐gene‐environmental causation of mental illness.

      ). In short, we have started to decipher the genetic code for SCZ, but many questions remain unanswered.
      Figure thumbnail gr1
      Figure 1History of genetic studies in schizophrenia spanning from pre-molecular genetics to current methodology, approaches, and findings, 1916: (118); 1966-1967-1971: (119–121); 1975: (122); 2002: (123); 2006: (124); 2008: (125); 2009: (11); 2011: (126); 2014: (12); 2017: (127).
      One of these questions is how genetic risk translates into the clinical manifestations of SCZ and its heterogeneous symptoms. GWAS approaches allow computing cumulative risk for a disorder by aggregating the effects of single nucleotide polymorphisms (SNPs) into a so-called polygenic risk score (PRS) (
      • Purcell S.M.
      • Wray N.R.
      • Stone J.L.
      • Visscher P.M.
      • O'Donovan M.C.
      • Sullivan P.F.
      • et al.
      Common polygenic variation contributes to risk of schizophrenia and bipolar disorder.
      ). However, low sensitivity and specificity limit the clinical utility of the PRS (
      • Purcell S.M.
      • Wray N.R.
      • Stone J.L.
      • Visscher P.M.
      • O'Donovan M.C.
      • Sullivan P.F.
      • et al.
      Common polygenic variation contributes to risk of schizophrenia and bipolar disorder.
      ,
      • Ripke S.
      • Neale B.M.
      • Corvin A.
      • Walters J.T.
      • Farh K.-H.
      • Holmans P.A.
      • et al.
      Biological insights from 108 schizophrenia-associated genetic loci.
      ,
      • Pardiñas A.F.
      • Holmans P.
      • Pocklington A.J.
      • Escott-Price V.
      • Ripke S.
      • Carrera N.
      • et al.
      Common schizophrenia alleles are enriched in mutation-intolerant genes and in regions under strong background selection.
      ). Critically, a high predicted genetic risk for SCZ across individuals does not necessarily imply a homogeneously dysregulated biology. Two individuals with equal scores likely have different risk alleles, meaning that disorder-relevant biological pathways may be affected differently in individuals with similar PRS (
      • Martin A.R.
      • Daly M.J.
      • Robinson E.B.
      • Hyman S.E.
      • Neale B.M.
      Predicting Polygenic Risk of Psychiatric Disorders.
      ). In other words, if risk is not unidimensional but relies on multiple intertwined or alternative pathways, different aetiologies may bring about different symptoms (
      • Docherty A.R.
      • Bigdeli T.B.
      • Edwards A.C.
      • Bacanu S.
      • Lee D.
      • Neale M.C.
      • et al.
      Genome-wide gene pathway analysis of psychotic illness symptom dimensions based on a new schizophrenia-specific model of the OPCRIT.
      ,
      • Goff D.C.
      • Tsai G.
      • Manoach D.S.
      • Coyle J.T.
      Dose-Finding Trial of D-Cycloserine Added to Neuroleptics for Negative SYmptoms in Schizophrenia.
      ,
      • Goff D.C.
      • Tsai G.
      • Levitt J.
      • Amico E.
      • Manoach D.
      • Schoenfeld D.A.
      • et al.
      A Placebo-Controlled Trial of D-Cycloserine Added to Conventional Neuroleptics in Patients With Schizophrenia.
      ). This idea lingers in recent debates about defining a psychosis continuum akin to broad spectrum phenotypes like autism spectrum disorder introduced in the DSM-V (
      • Guloksuz S.
      • van Os J.
      The slow death of the concept of schizophrenia and the painful birth of the psychosis spectrum.
      ).
      Although the genetic architecture of SCZ has not been ultimately defined, evidence suggests it involves many genes and is aggregated into pathways, reflecting a certain degree of risk coherence in the framework put forward by Kendler (2013) (
      • Kendler K.S.
      What psychiatric genetics has taught us about the nature of psychiatric illness and what is left to learn.
      ). For instance, a model alternative to the polygenic theory argues that all genetic variants bear a non-zero effect on SCZ risk (omnigenic model: Liu et al., (2019) (
      • Liu X.
      • Li Y.I.
      • Pritchard J.K.
      Trans Effects on Gene Expression Can Drive Omnigenic Inheritance.
      ). However, even in an omnigenic context, gene regulators such as miR-137 result significantly enriched for genetic risk, suggesting that there is coherence in gene sets associated with potential master regulators (
      • Rammos A.
      • Gonzalez L.A.N.
      • Weinberger D.R.
      • Mitchell K.J.
      • Nicodemus K.K.
      The role of polygenic risk score gene-set analysis in the context of the omnigenic model of schizophrenia.
      )).
      To the extent that genetic risk for SCZ has some coherence, its complexity could be reduced by investigating the potential biological function of convergence pathways. An account of how genetic risk translates into psychiatric disorders that aptly considers the heterogeneity between patients and the heterogeneity of the biological pathways involved is the goal of precision psychiatry (
      • Liu X.
      • Li Y.I.
      • Pritchard J.K.
      Trans Effects on Gene Expression Can Drive Omnigenic Inheritance.
      ). This review will refer to the process of attributing genetic variants to molecular pathways as “genetic risk parsing”. We aim to assess whether genetic risk parsing has impacted our knowledge of (i) neurobiological underpinnings of SCZ, and (ii) potentially actionable insights for the clinic and drug discovery.
      Figure_1_about_here

      1. Approaches to parse genetic risk into pathways.

      As a methodological note of caution, parsing rests on the assumption that genetic variants effectively map onto genes. However, methodological choices to associate genes with variants impact the inferences drawn from genetic risk parsing (see Supplemental Information for details). Here, we consider three approaches that have been used to prioritize putative risk genes converging into relevant biological pathways, graphically summarized in Figure 2:
      Figure thumbnail gr2
      Figure 2Schematic figure of a study design parsing genetic variants. A) Genes can be related to meaningful pathways in multiple ways (gene ontology, co-expression, targetome; see chapters 2 and 3 for more details). B) Genetic variants are paired to these pathway-specific genes with a method of choice (e.g., MAGMA, PrediXcan: Supplemental Information). C) Parsed variants associated with a disease (risk SNPs) can be used to compute a genetic risk score; parsed variants associated with gene expression can be used to compute a score to predict pathway-specific gene expression. D) These scores can now be used to associate them with intermediate phenotypes or use them for predicting treatment outcomes.

      a.) Reference-based approaches

      Reference-based approaches (e.g., gene ontologies, reactome or the Kyoto Encyclopedia of Genes and Genomes [KEGG]) link a gene product such as a protein with biological processes, molecular functions, and cellular components (
      • Hill D.P.
      • Smith B.
      • McAndrews-Hill M.S.
      • Blake J.A.
      Gene Ontology annotations: what they mean and where they come from.
      ). SNPs related to the genes linked with biological processes of interest, e.g., glutamatergic signaling, can be prioritized in creating pathway-specific PRSs (
      • Rampino A.
      • Taurisano P.
      • Fanelli G.
      • Attrotto M.
      • Torretta S.
      • Antonucci L.A.
      • et al.
      A Polygenic Risk Score of glutamatergic SNPs associated with schizophrenia predicts attentional behavior and related brain activity in healthy humans.
      ). Here, we focus on gene ontologies, cell specificity and KEGG as the most common reference-based parsing approach in SCZ research so far.

      b.) Regulome

      Another approach considers that most of the risk variants identified by GWASs are located in non-coding DNA sequences that have a critical role in controlling gene expression, i.e., regulatory elements (REs) (

      Ripke S, Walters JTR, O’Donovan MC (2020): Mapping genomic loci prioritises genes and implicates synaptic biology in schizophrenia.

      ,
      • Maurano M.T.
      • Humbert R.
      • Rynes E.
      • Thurman R.E.
      • Haugen E.
      • Wang H.
      • et al.
      Systematic localization of common disease-associated variation in regulatory DNA.
      ). To date, the most prominent effect of gene regulation for SCZ risk was shown for microRNAs, e.g., miR-137 (
      • Hill D.P.
      • Smith B.
      • McAndrews-Hill M.S.
      • Blake J.A.
      Gene Ontology annotations: what they mean and where they come from.
      ,
      • Potkin S.G.
      • Macciardi F.
      • Guffanti G.
      • Fallon J.H.
      • Wang Q.
      • Turner J.A.
      • et al.
      Identifying gene regulatory networks in schizophrenia.
      ,
      • Mothersill O.
      • Morris D.W.
      • Kelly S.
      • Rose E.J.
      • Fahey C.
      • O'Brien C.
      • et al.
      Effects of MIR137 on fronto-amygdala functional connectivity.
      ,
      • Hauberg M.E.
      • Roussos P.
      • Grove J.
      • Børglum A.D.
      • Mattheisen M.
      Analyzing the Role of MicroRNAs in Schizophrenia in the Context of Common Genetic Risk Variants.
      ) and some transcription factors (TF), such as TCF4, SP1, ZNF804A (
      • Forrest M.P.
      • Hill M.J.
      • Kavanagh D.H.
      • Tansey K.E.
      • Waite A.J.
      • Blake D.J.
      The Psychiatric Risk Gene Transcription Factor 4 (TCF4) Regulates Neurodevelopmental Pathways Associated With Schizophrenia, Autism, and Intellectual Disability.
      ,
      • Ben-Shachar D.
      • Karry R.
      Sp1 expression is disrupted in schizophrenia; a possible mechanism for the abnormal expression of mitochondrial complex I genes, NDUFV1 and NDUFV2.
      ,
      • Chang H.
      • Xiao X.
      • Li M.
      The schizophrenia risk gene ZNF804A: clinical associations, biological mechanisms and neuronal functions.
      ). Parsed PRSs can be generated based on the definition of target genes of master regulators related to SCZ risk (
      • Rammos A.
      • Gonzalez L.A.N.
      • Weinberger D.R.
      • Mitchell K.J.
      • Nicodemus K.K.
      The role of polygenic risk score gene-set analysis in the context of the omnigenic model of schizophrenia.
      ,
      • Cosgrove D.
      • Harold D.
      • Mothersill O.
      • Anney R.
      • Hill M.J.
      • Bray N.J.
      • et al.
      MiR-137-derived polygenic risk: effects on cognitive performance in patients with schizophrenia and controls.
      ,
      • Amariuta T.
      • Ishigaki K.
      • Sugishita H.
      • Ohta T.
      • Koido M.
      • Dey K.K.
      • et al.
      Improving the trans-ancestry portability of polygenic risk scores by prioritizing variants in predicted cell-type-specific regulatory elements.
      ).

      c.) Co-expression

      The expression of genes is correlated to orchestrate cellular responses to stimuli (
      • Gaiteri C.
      • Ding Y.
      • French B.
      • Tseng G.C.
      • Sibille E.
      Beyond modules and hubs: the potential of gene coexpression networks for investigating molecular mechanisms of complex brain disorders.
      ,
      • Parikshak N.N.
      • Gandal M.J.
      • Geschwind D.H.
      Systems biology and gene networks in neurodevelopmental and neurodegenerative disorders.
      ). Shared transcription, maturational, and degradation processes along with brain function are thought to underlie the covariation of expression between genes, i.e., co-expression (
      • Fromer M.
      • Roussos P.
      • Sieberts S.K.
      • Johnson J.S.
      • Kavanagh D.H.
      • Perumal T.M.
      • et al.
      Gene expression elucidates functional impact of polygenic risk for schizophrenia.
      ). Popular approaches in this field identify clusters – so-called modules – of genes with highly correlated gene expression. Co-expression of gene sets within these modules can be summarized into one score by their first principal component, the so-called module eigengene, that subsequently can be associated with traits of interest (
      • Gandal M.J.
      • Haney J.R.
      • Parikshak N.N.
      • Leppa V.
      • Ramaswami G.
      • Hartl C.
      • et al.
      Shared molecular neuropathology across major psychiatric disorders parallels polygenic overlap.
      ).
      Figure_2_about_here

      2. Neuroimaging and clinical translation of parsed risk

      a) Reference-based approaches

      Reference-based approaches can be implemented in a data-driven fashion by testing all possible pathways involved or selecting pathways of interest in a hypothesis-driven manner. Several reports have used traditional etiopathogenic hypotheses on SCZ risk to parse its biology, e.g., pointing to dopaminergic and glutamatergic transmission. For example, Wang et al., (2018) computed a PRS including only SCZ risk variants (
      • Ripke S.
      • Neale B.M.
      • Corvin A.
      • Walters J.T.
      • Farh K.-H.
      • Holmans P.A.
      • et al.
      Biological insights from 108 schizophrenia-associated genetic loci.
      ) that mapped to a set of dopamine-related genes (
      • Wang C.
      • Liu B.
      • Zhang X.
      • Cui Y.
      • Yu C.
      • Jiang T.
      Multilocus genetic profile in dopaminergic pathway modulates the striatum and working memory.
      ). A higher dopaminergic PRS score was associated with poorer working memory performance in 323 healthy individuals (
      • Wang C.
      • Liu B.
      • Zhang X.
      • Cui Y.
      • Yu C.
      • Jiang T.
      Multilocus genetic profile in dopaminergic pathway modulates the striatum and working memory.
      ) in parallel with striatal dysconnectivity during an fMRI working memory task. Similarly, building upon previous associations of glutamatergic neurotransmission with attention (
      • Gallinat J.
      • Götz T.
      • Kalus P.
      • Bajbouj M.
      • Sander T.
      • Winterer G.
      Genetic Variations of the NR3A Subunit of the NMDA Receptor Modulate Prefrontal Cerebral Activity in Humans.
      ) – a cognitive domain reproducibly associated with SCZ (
      • Nuechterlein K.H.
      • Barch D.M.
      • Gold J.M.
      • Goldberg T.E.
      • Green M.F.
      • Heaton R.K.
      Identification of separable cognitive factors in schizophrenia.
      ) – Rampino et al., (2017) calculated a glutamatergic PRS by selecting SCZ risk variants (
      • Ripke S.
      • Neale B.M.
      • Corvin A.
      • Walters J.T.
      • Farh K.-H.
      • Holmans P.A.
      • et al.
      Biological insights from 108 schizophrenia-associated genetic loci.
      ) in the proximity of genes coding for glutamatergic synapsis proteins (
      • Rampino A.
      • Taurisano P.
      • Fanelli G.
      • Attrotto M.
      • Torretta S.
      • Antonucci L.A.
      • et al.
      A Polygenic Risk Score of glutamatergic SNPs associated with schizophrenia predicts attentional behavior and related brain activity in healthy humans.
      ). The glutamatergic but not the non-parsed whole-genome PRS was associated with brain activation during the visual attention control task employed (
      • Blasi G.
      • Mattay V.S.
      • Bertolino A.
      • Elvevåg B.
      • Callicott J.H.
      • Das S.
      • et al.
      Effect of catechol-O-methyltransferase val158met genotype on attentional control.
      ). These findings on a single ontology pathway suggest that selecting SNPs involved in specific molecular processes may help link specific risk variants with specific deficits characterizing patients with SCZ. Adopting a set of five pathways previously linked with SCZ, Barbu et al., (2022) successfully associated three of them with structural brain features (

      Barbu MC, Thng G, Adams MJ, Marwick K, Grant SGN, McIntosh AM, et al. (2022): Pathway-based polygenic risk scores for schizophrenia and associations with clinical and neuroimaging phenotypes in UK Biobank.

      ). However, these hypothesis-driven approaches suffer from restricting the search space to select pathways derived from databases created from different tissues and species.
      Other studies defined pathways based on cell specificity. Corley et al., (2021) computed a PRS selectively on SCZ risk alleles previously linked to microglia, neurons, and astroglia (
      • Pardiñas A.F.
      • Holmans P.
      • Pocklington A.J.
      • Escott-Price V.
      • Ripke S.
      • Carrera N.
      • et al.
      Common schizophrenia alleles are enriched in mutation-intolerant genes and in regions under strong background selection.
      ). However, the reproducible PRS associations with cognitive performance showed poor specificity for the PRSs parsed according to cell types (
      • Corley E.
      • Holleran L.
      • Fahey L.
      • Corvin A.
      • Morris D.W.
      • Donohoe G.
      Microglial-expressed genetic risk variants, cognitive function and brain volume in patients with schizophrenia and healthy controls.
      ). Another report found that a PRS based on SCZ-associated SNPs that are part of cell adhesion molecule pathways correlated with memory and attention in patients with psychosis (
      • Hargreaves A.
      • Anney R.
      • O'Dushlaine C.
      • Nicodemus K.K.
      • Gill M.
      • Corvin A.
      • et al.
      The one and the many: effects of the cell adhesion molecule pathway on neuropsychological function in psychosis.
      ). Results in attentional performance, however, appeared driven by a single SNP. Using a similar parsing approach, Di Biase et al., (2022) examined 1,927 patients and controls and reported that the pattern of regional cortical thickness deviations in patients reflected a genetic predisposition mainly weighing on the cell types in which genes proximal to such genetic variants were expressed (
      • Di Biase M.A.
      • Geaghan M.P.
      • Reay W.R.
      • Seidlitz J.
      • Weickert C.S.
      • Pébay A.
      • et al.
      Cell type-specific manifestations of cortical thickness heterogeneity in schizophrenia.
      ). These studies suggest that linking multiple biological layers based on biological pathways of risk might contribute to explaining heterogeneity between brain abnormalities seen in SCZ in terms of clinical subtyping; however, the specificity and generalizability of these approaches seem to require further support.
      A key question is whether adding biological priors to our predictions actually increases prediction accuracy. A data-driven study leveraged the large data pool from the PsychENCODE dataset (
      • Akbarian S.
      • Liu C.
      • Knowles J.A.
      • Vaccarino F.M.
      • Farnham P.J.
      • Crawford G.E.
      • et al.
      The PsychENCODE project.
      ) to integrate biological priors into genetic risk-based predictions (

      Wang D, Liu S, Warrell J, Won H, Shi X, Navarro FCP, et al. (2018): Comprehensive functional genomic resource and integrative model for the human brain. Science (New York, N.Y.) 362.

      ). They fed cell-specific gene expression signatures of risk, chromatin accessibility data, and eQTLs into a deep learning model. Integrating biological priors into genetic risk-based predictions enhanced the accuracy of case/control classification about six-fold compared to a “benchmark” PRS (

      Wang D, Liu S, Warrell J, Won H, Shi X, Navarro FCP, et al. (2018): Comprehensive functional genomic resource and integrative model for the human brain. Science (New York, N.Y.) 362.

      ). Even using imputed gene expression rather than actual gene expression achieved better predictions than those based solely on GWAS, suggesting that these algorithms could be employed for living patients for whom molecular brain data are not accessible (

      Ghosal S, Chen Q, Pergola G, Goldman AL, Ulrich W, Weinberger DR, et al. (2021): A Biologically Interpretable Graph Convolutional Network to Link Genetic Risk Pathways and Neuroimaging Markers of Disease.

      ).
      As shown by Wang et al., (2018), parsing is not limited to a single labeling: a two-stage machine learning approach called “biologically informed machine learning (BioMM)” (

      Chen J, Schwarz E (01/12/2017): BioMM: Biologically-informed Multi-stage Machine learning for identification of epigenetic fingerprints.

      ,
      • Chen J.
      • Zang Z.
      • Braun U.
      • Schwarz K.
      • Harneit A.
      • Kremer T.
      • et al.
      Association of a Reproducible Epigenetic Risk Profile for Schizophrenia With Brain Methylation and Function.
      ) parsed methylation data into different pathways based on gene ontology categories to classify between HCs and SCZ patients (

      Wang D, Liu S, Warrell J, Won H, Shi X, Navarro FCP, et al. (2018): Comprehensive functional genomic resource and integrative model for the human brain. Science (New York, N.Y.) 362.

      ). The polymethylation score they generated, representing an epigenetic signature of SCZ, was significantly associated with DLPFC-hippocampal connectivity during working memory, representing a candidate intermediate phenotype of SCZ. Intermediate phenotypes are heritable neurobiological measures associated with risk for a heritable disorder (
      • Gottesman II,
      • Gould T.D.
      The Endophenotype Concept in Psychiatry: Etymology and Strategic Intentions.
      ,
      • Meyer-Lindenberg A.
      • Weinberger D.
      Intermediate phenotypes and genetic mechanisms of psychiatric disorders.
      ,
      • Bertolino A.
      • Blasi G.
      The genetics of schizophrenia.
      ). When using the BioMM model for patient classification, Chen et al., (2020) (
      • Chen J.
      • Zang Z.
      • Braun U.
      • Schwarz K.
      • Harneit A.
      • Kremer T.
      • et al.
      Association of a Reproducible Epigenetic Risk Profile for Schizophrenia With Brain Methylation and Function.
      ) reported areas under the curves between 0.69 and 0.78, hence above the threshold of 0.7 sometimes deemed clinically useful (
      • Mandrekar J.N.
      Receiver operating characteristic curve in diagnostic test assessment.
      ). Although this score is not based on genetic sequence, it shows once again that including prior biological knowledge on risk parsing can effectively boost predictions about SCZ.
      However, reference-based “genetic risk parsing” also returned negative findings (
      • Janouschek H.
      • Eickhoff C.R.
      • Mühleisen T.W.
      • Eickhoff S.B.
      • Nickl-Jockschat T.
      Using coordinate-based meta-analyses to explore structural imaging genetics.
      ). Besides mixed results, the inherent incompleteness of the annotations, which depend on prior biological knowledge, and the methods used to compile the lists limit the use of gene ontologies as parsing criteria (
      • Janouschek H.
      • Eickhoff C.R.
      • Mühleisen T.W.
      • Eickhoff S.B.
      • Nickl-Jockschat T.
      Using coordinate-based meta-analyses to explore structural imaging genetics.
      ). These labels change over time and vary across genomic regions (
      • Gillis J.
      • Pavlidis P.
      Assessing identity, redundancy and confounds in Gene Ontology annotations over time.
      ). Gene ontology pathways reflect similar protein functions and not evidence-based biological relationships between the genes (
      • Gillis J.
      • Pavlidis P.
      Characterizing the state of the art in the computational assignment of gene function: lessons from the first critical assessment of functional annotation (CAFA).
      ). These inconsistencies may be a cause of replication failures (
      • Bleazard T.
      • Lamb J.A.
      • Griffiths-Jones S.
      Bias in microRNA functional enrichment analysis.
      ). The circumstance that gene ontology catalogs are compiled based on experimental evidence from different tissues and even species hinders their applications to human psychiatric disorders.

      b) Regulome

      Previous studies have shown that many associations identified by GWASs were attributable to variants located in REs (
      • Ripke S.
      • Neale B.M.
      • Corvin A.
      • Walters J.T.
      • Farh K.-H.
      • Holmans P.A.
      • et al.
      Biological insights from 108 schizophrenia-associated genetic loci.
      ,

      Ripke S, Walters JTR, O’Donovan MC (2020): Mapping genomic loci prioritises genes and implicates synaptic biology in schizophrenia.

      ,
      • Maurano M.T.
      • Humbert R.
      • Rynes E.
      • Thurman R.E.
      • Haugen E.
      • Wang H.
      • et al.
      Systematic localization of common disease-associated variation in regulatory DNA.
      ). The relationship between REs and risk variants has been used to identify causal or functional variants for complex diseases (
      • Bond G.L.
      • Hu W.
      • Bond E.E.
      • Robins H.
      • Lutzker S.G.
      • Arva N.C.
      • et al.
      A single nucleotide polymorphism in the MDM2 promoter attenuates the p53 tumor suppressor pathway and accelerates tumor formation in humans.
      ). REs usually contain multiple binding sites (i.e., DNA motifs that TFs can recognize); genetic variation in REs affects gene expression by altering the binding affinity of TFs in tissue- and cell-type specific fashion (
      • Joshi M.
      • Kapopoulou A.
      • Laurent S.
      Impact of Genetic Variation in Gene Regulatory Sequences: A Population Genomics Perspective.
      ).
      Thus, regulome parsing studies have generally focused on restricting genetic risk estimation to the target genes of specific master regulators. For example, Wright et al., (2016) investigated the interaction between a miR-137 variant (rs1625579) and variants in four experimentally validated miR-137 target genes (
      • Wright C.
      • Gupta C.N.
      • Chen J.
      • Patel V.
      • Calhoun V.D.
      • Ehrlich S.
      • et al.
      Polymorphisms in MIR137HG and microRNA-137-regulated genes influence gray matter structure in schizophrenia.
      ). However, they did not report reproducible genetic effects. Instead, Cosgrove et al., (2017) has reported an association between working memory-related brain activity and genetic variants proximal to genes potentially regulated by miR-137 (
      • Cosgrove D.
      • Harold D.
      • Mothersill O.
      • Anney R.
      • Hill M.J.
      • Bray N.J.
      • et al.
      MiR-137-derived polygenic risk: effects on cognitive performance in patients with schizophrenia and controls.
      ) according to Hill et al., (2008) (
      • Hill D.P.
      • Smith B.
      • McAndrews-Hill M.S.
      • Blake J.A.
      Gene Ontology annotations: what they mean and where they come from.
      ). This gene set generated PRSs based on risk SNPs proximal to target genes (
      • Ripke S.
      • Neale B.M.
      • Corvin A.
      • Walters J.T.
      • Farh K.-H.
      • Holmans P.A.
      • et al.
      Biological insights from 108 schizophrenia-associated genetic loci.
      ). The PRS in this gene set was associated with lower declarative memory performance and nominally associated with poorer working memory and general cognitive ability in healthy controls and patients with SCZ. The effect of the miR-137 selective PRS on working memory brain activity has been preliminarily replicated in a larger sample (

      Pergola G, Rampino A, Di Carlo P, Marakhovskaia A, Quarto T, Fazio L, et al. (2020): A miR-137-related biological pathway of risk for Schizophrenia is associated with human brain emotion processing.

      ). A further study investigating the effect of this PRS on brain structure returned only weak results, suggesting there is some specificity to phenotypic associations obtained via RE risk parsing (
      • Cosgrove D.
      • Mothersill D.O.
      • Whitton L.
      • Harold D.
      • Kelly S.
      • Holleran L.
      • et al.
      Effects of MiR-137 genetic risk score on brain volume and cortical measures in patients with schizophrenia and controls.
      ): parsed PRSs capture different phenomena of the disorder, as can be expected in a scenario of coherent genetic risk converging into biological pathways. Additionally, the miR-137 PRS explained disproportionally larger variance of SCZ risk compared with other parsing approaches and whole-genome PRS considering the number of SNPs (
      • Yao Y.
      • Guo W.
      • Zhang S.
      • Yu H.
      • Yan H.
      • Zhang H.
      • et al.
      Cell type-specific and cross-population polygenic risk score analyses of MIR137 gene pathway in schizophrenia.
      ). Regulomic approaches seem promising in terms of biological plausibility and specificity (
      • Sey N.Y.A.
      • Hu B.
      • Mah W.
      • Fauni H.
      • McAfee J.C.
      • Rajarajan P.
      • et al.
      A computational tool (H-MAGMA) for improved prediction of brain-disorder risk genes by incorporating brain chromatin interaction profiles.
      ,

      Casella AM, Colantuoni C, Ament SA (2021): Regulome-wide association study identifies enhancer properties associated with risk for schizophrenia.

      ), although they have some limitations shared with reference-based approaches. For example, most target attributions are generated via bioinformatic predictions. Even for one of the most plausible regulators of SCZ risk genes, i.e., miR-137, only about 50 target genes have been experimentally validated out of the thousands predicted (
      • Mahmoudi E.
      • Cairns M.J.
      MiR-137: an important player in neural development and neoplastic transformation.
      ).

      c) Co-expression

      An essential premise to the idea that genetic risk for a psychiatric disorder is coherent is that, even if variants are spread across the entire genome, their functional correlates converge into molecular or systems-level processes (
      • Harrison P.J.
      • Weinberger D.R.
      Schizophrenia genes, gene expression, and neuropathology: on the matter of their convergence.
      ). Gene expression assays have provided evidence that SCZ risk genes converge into some co-expression modules more than others, i.e., modules overrepresenting putative SCZ risk genes (
      • Fromer M.
      • Roussos P.
      • Sieberts S.K.
      • Johnson J.S.
      • Kavanagh D.H.
      • Perumal T.M.
      • et al.
      Gene expression elucidates functional impact of polygenic risk for schizophrenia.
      ,
      • Roussos P.
      • Katsel P.
      • Davis K.L.
      • Siever L.J.
      • Haroutunian V.
      A system-level transcriptomic analysis of schizophrenia using postmortem brain tissue samples.
      ,
      • Roussos P.
      • Guennewig B.
      • Kaczorowski D.C.
      • Barry G.
      • Brennand K.J.
      Activity-Dependent Changes in Gene Expression in Schizophrenia Human-Induced Pluripotent Stem Cell Neurons.
      ,

      Li M, Santpere G, Imamura Kawasawa Y, Evgrafov OV, Gulden FO, Pochareddy S, et al. (2018): Integrative functional genomic analysis of human brain development and neuropsychiatric risks. Science (New York, N.Y.) 362.

      ,
      • Walker R.L.
      • Ramaswami G.
      • Hartl C.
      • Mancuso N.
      • Gandal M.J.
      • La Torre-Ubieta L de
      • et al.
      Genetic Control of Expression and Splicing in Developing Human Brain Informs Disease Mechanisms.
      ,
      • Werling D.M.
      • Pochareddy S.
      • Choi J.
      • An J.-Y.
      • Sheppard B.
      • Peng M.
      • et al.
      Whole-Genome and RNA Sequencing Reveal Variation and Transcriptomic Coordination in the Developing Human Prefrontal Cortex.
      ,
      • Hartl C.L.
      • Ramaswami G.
      • Pembroke W.G.
      • Muller S.
      • Pintacuda G.
      • Saha A.
      • et al.
      Coexpression network architecture reveals the brain-wide and multiregional basis of disease susceptibility.
      ,
      • Jia P.
      • Manuel A.M.
      • Fernandes B.S.
      • Dai Y.
      • Zhao Z.
      Distinct effect of prenatal and postnatal brain expression across 20 brain disorders and anthropometric social traits: a systematic study of spatiotemporal modularity.
      ,

      Gandal MJ, Zhang P, Hadjimichael E, Walker RL, Chen C, Liu S, et al. (2018): Transcriptome-wide isoform-level dysregulation in ASD, schizophrenia, and bipolar disorder. Science (New York, N.Y.) 362.

      ,
      • Pergola G.
      • Di Carlo P.
      • Jaffe A.E.
      • Papalino M.
      • Chen Q.
      • Hyde T.M.
      • et al.
      Prefrontal Coexpression of Schizophrenia Risk Genes Is Associated With Treatment Response in Patients.
      ,
      • Radulescu E.
      • Jaffe A.E.
      • Straub R.E.
      • Chen Q.
      • Shin J.H.
      • Hyde T.M.
      • et al.
      Identification and prioritization of gene sets associated with schizophrenia risk by co-expression network analysis in human brain.
      ). Interestingly, studies defining putative SCZ risk genes based on rare genetic variation, such as de novo mutations, have also found coherent gene co-expression in the prefrontal cortex (
      • Fromer M.
      • Roussos P.
      • Sieberts S.K.
      • Johnson J.S.
      • Kavanagh D.H.
      • Perumal T.M.
      • et al.
      Gene expression elucidates functional impact of polygenic risk for schizophrenia.
      ,
      • Gulsuner S.
      • Walsh T.
      • Watts A.C.
      • Lee M.K.
      • Thornton A.M.
      • Casadei S.
      • et al.
      Spatial and temporal mapping of de novo mutations in schizophrenia to a fetal prefrontal cortical network.
      ,
      • Wang Q.
      • Li M.
      • Yang Z.
      • Hu X.
      • Wu H.-M.
      • Ni P.
      • et al.
      Increased co-expression of genes harboring the damaging de novo mutations in Chinese schizophrenic patients during prenatal development.
      ). Co-expression gene sets enriched for risk genes for SCZ are also enriched for neuronal markers, axon development, synapse function, and hemophilic cell adhesion (
      • Fromer M.
      • Roussos P.
      • Sieberts S.K.
      • Johnson J.S.
      • Kavanagh D.H.
      • Perumal T.M.
      • et al.
      Gene expression elucidates functional impact of polygenic risk for schizophrenia.
      ,
      • Gandal M.J.
      • Haney J.R.
      • Parikshak N.N.
      • Leppa V.
      • Ramaswami G.
      • Hartl C.
      • et al.
      Shared molecular neuropathology across major psychiatric disorders parallels polygenic overlap.
      ,
      • Roussos P.
      • Katsel P.
      • Davis K.L.
      • Siever L.J.
      • Haroutunian V.
      A system-level transcriptomic analysis of schizophrenia using postmortem brain tissue samples.
      ,

      Gandal MJ, Zhang P, Hadjimichael E, Walker RL, Chen C, Liu S, et al. (2018): Transcriptome-wide isoform-level dysregulation in ASD, schizophrenia, and bipolar disorder. Science (New York, N.Y.) 362.

      ,
      • Pergola G.
      • Di Carlo P.
      • Jaffe A.E.
      • Papalino M.
      • Chen Q.
      • Hyde T.M.
      • et al.
      Prefrontal Coexpression of Schizophrenia Risk Genes Is Associated With Treatment Response in Patients.
      ,
      • Radulescu E.
      • Jaffe A.E.
      • Straub R.E.
      • Chen Q.
      • Shin J.H.
      • Hyde T.M.
      • et al.
      Identification and prioritization of gene sets associated with schizophrenia risk by co-expression network analysis in human brain.
      ,
      • Gulsuner S.
      • Walsh T.
      • Watts A.C.
      • Lee M.K.
      • Thornton A.M.
      • Casadei S.
      • et al.
      Spatial and temporal mapping of de novo mutations in schizophrenia to a fetal prefrontal cortical network.
      ). These findings suggest that at least some of the convergence of genetic risk into synaptic biology is apparent at the level of gene co-expression (

      Ripke S, Walters JTR, O’Donovan MC (2020): Mapping genomic loci prioritises genes and implicates synaptic biology in schizophrenia.

      ,
      • Kim M.
      • Haney J.R.
      • Zhang P.
      • Hernandez L.M.
      • Wang L.-K.
      • Perez-Cano L.
      • et al.
      Brain gene co-expression networks link complement signaling with convergent synaptic pathology in schizophrenia.
      ). Finally, some studies reported co-expression associations with SCZ-PRS (
      • Wang C.
      • Liu B.
      • Zhang X.
      • Cui Y.
      • Yu C.
      • Jiang T.
      Multilocus genetic profile in dopaminergic pathway modulates the striatum and working memory.
      ,
      • Radulescu E.
      • Jaffe A.E.
      • Straub R.E.
      • Chen Q.
      • Shin J.H.
      • Hyde T.M.
      • et al.
      Identification and prioritization of gene sets associated with schizophrenia risk by co-expression network analysis in human brain.
      ,
      • Hauberg M.E.
      • Fullard J.F.
      • Zhu L.
      • Cohain A.T.
      • Giambartolomei C.
      • Misir R.
      • et al.
      Differential activity of transcribed enhancers in the prefrontal cortex of 537 cases with schizophrenia and controls.
      ). In summary, putative SCZ risk genes tend to be co-expressed above chance, although it is unknown whether this concurrence reflects convergent biological processes.
      As reviewed above, integrating genetic information with biological priors is instrumental in explaining genetic risk and may improve case-control discrimination along with the genotypes (
      • Wang C.
      • Liu B.
      • Zhang X.
      • Cui Y.
      • Yu C.
      • Jiang T.
      Multilocus genetic profile in dopaminergic pathway modulates the striatum and working memory.
      ,
      • Li Z.
      • Li X.
      • Jin M.
      • Liu Y.
      • He Y.
      • Jia N.
      • et al.
      Identification of potential biomarkers and their correlation with immune infiltration cells in schizophrenia using combinative bioinformatics strategy.
      ,
      • Hess J.L.
      • Tylee D.S.
      • Barve R.
      • Jong S de
      • Ophoff R.A.
      • Kumarasinghe N.
      • et al.
      Transcriptome-wide mega-analyses reveal joint dysregulation of immunologic genes and transcription regulators in brain and blood in schizophrenia.
      ,
      • Chen J.
      • Cao H.
      • Kaufmann T.
      • Westlye L.T.
      • Tost H.
      • Meyer-Lindenberg A.
      • et al.
      Identification of Reproducible BCL11A Alterations in Schizophrenia Through Individual-Level Prediction of Coexpression.
      ). On this basis, genetic variants associated with different co-expression gene sets could contribute to explaining genetic risk, phenotypes associated with it, and clinical heterogeneity. The idea that genetic variants might be associated with genes in correlated networks rests on the notion that trans-eQTLs may be associated with gene expression, although the genes are not in their immediate proximity (see Supplemental Information for details). In general, trans-eQTL effects are considered small compared to cis-eQTLs (
      • Gamazon E.R.
      • Wheeler H.E.
      • Shah K.P.
      • Mozaffari S.V.
      • Aquino-Michaels K.
      • Carroll R.J.
      • et al.
      A gene-based association method for mapping traits using reference transcriptome data.
      ). Nevertheless, Hore et al. (
      • Hore V.
      • Viñuela A.
      • Buil A.
      • Knight J.
      • McCarthy M.I.
      • Small K.
      • et al.
      Tensor decomposition for multiple-tissue gene expression experiments.
      ) demonstrated that gene co-expression effectively identifies genome-wide trans-eQTLs associated with correlated gene sets. The SNPs that physically map to each gene within the set associated with the module eigengene have been called co-eQTLs (
      • Pergola G.
      • Di Carlo P.
      • D'Ambrosio E.
      • Gelao B.
      • Fazio L.
      • Papalino M.
      • et al.
      DRD2 co-expression network and a related polygenic index predict imaging, behavioral and clinical phenotypes linked to schizophrenia.
      ). The small effects of co-eQTLs can be combined into polygenic scores of imputed co-expression. These scores do not index risk for SCZ, but are transversal to healthy controls and other psychiatric diagnoses with the purpose of translating brain postmortem information into brain activity and clinical predictions in living subjects.
      This approach has been initially validated by leveraging prior knowledge on specific genes and their system-level correlates. For example, Pergola et al., (2017) identified a gene co-expression network in the postmortem DLPFC (
      • Pergola G.
      • Di Carlo P.
      • D'Ambrosio E.
      • Gelao B.
      • Fazio L.
      • Papalino M.
      • et al.
      DRD2 co-expression network and a related polygenic index predict imaging, behavioral and clinical phenotypes linked to schizophrenia.
      ). Focusing on the module including the DRD2 gene coding for the long isoform of the dopamine receptor D2, they generated a polygenic score predicting the simultaneous expression of DRD2 and its co-expression partners. In neurotypical controls undergoing an fMRI scan, greater predicted module expression was reproducibly associated with greater activation in a frontoparietal network and slower reaction times during 2-back working memory task performance. The same score originated from DLPFC is also associated to estimated striatal dopamine synthesis, suggesting the brain-wide relevance of the discovered genetic variants (
      • D'Ambrosio E.
      • Pergola G.
      • Pardiñas A.F.
      • Dahoun T.
      • Veronese M.
      • Sportelli L.
      • et al.
      A polygenic score indexing a DRD2-related co-expression network is associated with striatal dopamine function.
      ). A second genetic score predicting the co-expression of a module including the DRD1 gene, coding for the dopamine receptor D1 whose activity is opposite to D2 receptors, was reproducibly associated with fMRI prefrontal activity and behavioral performance in the opposite direction (
      • Fazio L.
      • Pergola G.
      • Papalino M.
      • Di Carlo P.
      • Monda A.
      • Gelao B.
      • et al.
      Transcriptomic context of DRD1 is associated with prefrontal activity and behavior during working memory.
      ). Investigating both scores in the context of brain connectivity revealed once again opposite effects of the D1 and D2 co-expression scores (
      • Braun U.
      • Harneit A.
      • Pergola G.
      • Menara T.
      • Schäfer A.
      • Betzel R.F.
      • et al.
      Brain network dynamics during working memory are modulated by dopamine and diminished in schizophrenia.
      ). At a clinical level, Pergola et al., (2017) also reported an association with treatment response to antipsychotics in two independent cohorts of patients with SCZ (
      • Pergola G.
      • Di Carlo P.
      • D'Ambrosio E.
      • Gelao B.
      • Fazio L.
      • Papalino M.
      • et al.
      DRD2 co-expression network and a related polygenic index predict imaging, behavioral and clinical phenotypes linked to schizophrenia.
      ). Using a data-driven approach for the identification of modules of interest, Pergola et al., (2019) reported that the co-expression of a module enriched for SCZ risk genes was reproducibly associated with short-term response to olanzapine in patients with SCZ – suggesting that the genetic variants within these modules could be instrumental in explaining part of the clinical heterogeneity (
      • Pergola G.
      • Di Carlo P.
      • Jaffe A.E.
      • Papalino M.
      • Chen Q.
      • Hyde T.M.
      • et al.
      Prefrontal Coexpression of Schizophrenia Risk Genes Is Associated With Treatment Response in Patients.
      ). Unlike a conventional GWAS-based PRS without biological priors, this approach can also test non-linear relationships between alleles associated with gene expression and brain physiology. For example, Selvaggi et al., (2019) found a reproducible quadratic association between predicted DRD2 module expression and brain activity during working memory that was reversed by administering the D2-targeting drug bromocriptine (

      Selvaggi P, Pergola G, Gelao B, Di Carlo P, Nettis MA, Amico G, et al. (2019): Genetic Variation of a DRD2 Co-expression Network is Associated with Changes in Prefrontal Function After D2 Receptors Stimulation. Cerebral cortex (New York, N.Y. : 1991) 29: 1162–1173.

      ). Another way of using networks consists of discovering genes in a key topological position. Rodríguez et al. (2022) clustered modules of co-expressed genes expressed in blood cells according to the association with symptomatology, level of functioning, and premorbid adjustment. The expression of modules’ hub genes of the two emerging clusters served as potential peripheral biomarkers of illness states or traits (
      • Rodríguez N.
      • Gassó P.
      • Martínez-Pinteño A.
      • Segura À.-G.
      • Mezquida G.
      • Moreno-Izco L.
      • et al.
      Gene co-expression architecture in peripheral blood in a cohort of remitted first-episode schizophrenia patients.
      ).
      Further studies have supported the specificity of co-eQTL associations with neuroimaging phenotypes. For example, a score more closely indexing the co-expression of the DRD2 gene with the CNR1 gene coding for the Cannabinoid Receptor 1, but not the original DRD2 score, interacted with cannabis exposure in its association with brain activity (
      • Monaco A.
      • Monda A.
      • Amoroso N.
      • Bertolino A.
      • Blasi G.
      • Di Carlo P.
      • et al.
      A complex network approach reveals a pivotal substructure of genes linked to schizophrenia.
      ,
      • Taurisano P.
      • Pergola G.
      • Monda A.
      • Antonucci L.A.
      • Di Carlo P.
      • Piarulli F.
      • et al.
      The interaction between cannabis use and a CB1-related polygenic co-expression index modulates dorsolateral prefrontal activity during working memory processing.
      ). Antonucci et al. (2019) studied the correlated expression of genes between the mediodorsal thalamic nucleus and the DLPFC to identify a thalamocortical module enriched for SCZ genetic risk. The genetic score predicting the co-expression of this module was associated with thalamocortical connectivity during attentional control (
      • Antonucci L.A.
      • Di Carlo P.
      • Passiatore R.
      • Papalino M.
      • Monda A.
      • Amoroso N.
      • et al.
      Thalamic connectivity measured with fMRI is associated with a polygenic index predicting thalamo-prefrontal gene co-expression.
      ). However, the same phenotype (previously proposed as a candidate intermediate phenotype, see (
      • Antonucci L.A.
      • Taurisano P.
      • Fazio L.
      • Gelao B.
      • Romano R.
      • Quarto T.
      • et al.
      Association of familial risk for schizophrenia with thalamic and medial prefrontal functional connectivity during attentional control.
      )) was not associated with DRD2 module predicted expression.
      In summary, translating co-expression into genetic predictors aligns with known neurophysiological effects, is associated with clinical profiles, and may potentially reveal novel biomarkers. A limitation of co-expression approaches is that, on the one hand, they do not imply co-transcription or co-regulation: for example, co-degradation likely contributes to the observed gene expression covariance (
      • Jaffe A.E.
      • Tao R.
      • Norris A.L.
      • Kealhofer M.
      • Nellore A.
      • Shin J.H.
      • et al.
      qSVA framework for RNA quality correction in differential expression analysis.
      ). Observed co-expression in bulk tissue stems from mRNA produced by different cell types, hence mixing cell type abundance with mRNA expression levels within cells (
      • Hartl C.L.
      • Ramaswami G.
      • Pembroke W.G.
      • Muller S.
      • Pintacuda G.
      • Saha A.
      • et al.
      Coexpression network architecture reveals the brain-wide and multiregional basis of disease susceptibility.
      ,
      • Parsana P.
      • Ruberman C.
      • Jaffe A.E.
      • Schatz M.C.
      • Battle A.
      • Leek J.T.
      Addressing confounding artifacts in reconstruction of gene co-expression networks.
      ). Limited definition of the biological properties underlying co-expression complicates the investigation of actionable biological mechanisms of disease.

      3. Lessons learned

      There is a considerable distance in terms of the biological organization layers involved between genetic risk and clinical phenotypes, i.e., transcriptomics, proteomics, cell physiology, system-level functioning, and behavior. Aggregating genetic risk in PRSs has many merits but does not contain specific biological information that can be readily translated. Here, we have critically reviewed studies that opted to parse genetic risk for SCZ. The published reports have been unsystematic regarding the methods employed. They often used select pathways, without always reporting negative controls, and procedures have been highly heterogeneous; overall, the literature is too sparse to reach definitive conclusions. There is initial evidence that biological priors improve prediction accuracy in diagnostic contexts, although biological labels lacking human and tissue specificity are probably unfit for understanding human brain physiology. Instead, biological information imputed from postmortem human brain tissue datasets may help further unravel the neurobiology of schizophrenia. For instance, the co-expression studies here reviewed are consistent with the known neurophysiology of cortical dopaminergic signaling. We consider it essential that any biological prior be consistent with what is already known about physiology and pathophysiology, without completely disregarding potential new ideas. For example, the postmortem and in vivo literature shows limited involvement of microglia in SCZ and, accordingly, the evaluation of the effect of genetic risk on microglia in this disorder lacks specificity. In pursuing co-expression approaches, the degree of time and cell resolution plays a paramount role, along with more biologically plausible ways of associating genetic variants with genes compared to the standard of genomic localization (Supplementary Information).
      The categorical diagnosis of SCZ is based on a collection of signs and symptoms that vary over time in the same patient (
      • Fisher A.J.
      • Medaglia J.D.
      • Jeronimus B.F.
      Lack of group-to-individual generalizability is a threat to human subjects research.
      ) and between patients (
      • Alnæs D.
      • Kaufmann T.
      • van der Meer D.
      • Córdova-Palomera A.
      • Rokicki J.
      • Moberget T.
      • et al.
      Brain Heterogeneity in Schizophrenia and Its Association With Polygenic Risk.
      ,
      • Bigdeli T.B.
      • Fanous A.H.
      • Li Y.
      • Rajeevan N.
      • Sayward F.
      • Genovese G.
      • et al.
      Genome-Wide Association Studies of Schizophrenia and Bipolar Disorder in a Diverse Cohort of US Veterans.
      ). Parsing genetic risk might help to identify sub-types of SCZ or, following other views, the SCZ spectrum disorders (
      • Guloksuz S.
      • van Os J.
      The slow death of the concept of schizophrenia and the painful birth of the psychosis spectrum.
      ). To the extent that different mechanisms of illness may be inherited based on different pathways, the genetics of SCZ may result as heterogeneous as its neurobiology and symptoms. Besides additive genetics and common variants, ultra-rare protein-altering genetic variation (
      • Genovese G.
      • Fromer M.
      • Stahl E.A.
      • Ruderfer D.M.
      • Chambert K.
      • Landén M.
      • et al.
      Increased burden of ultra-rare protein-altering variants among 4,877 individuals with schizophrenia.
      ,
      • Singh T.
      • Poterba T.
      • Curtis D.
      • Akil H.
      • Al Eissa M.
      • Barchas J.D.
      • et al.
      Rare coding variants in ten genes confer substantial risk for schizophrenia.
      ), and copy number variants (CNVs, (
      • Marshall C.R.
      • Howrigan D.P.
      • Merico D.
      • Thiruvahindrapuram B.
      • Wu W.
      • Greer D.S.
      • et al.
      Contribution of copy number variants to schizophrenia from a genome-wide study of 41,321 subjects.
      )) such as those found in the 16p11.2 (
      • Sønderby I.E.
      • Gústafsson Ó.
      • Doan N.T.
      • Hibar D.P.
      • Martin-Brevet S.
      • Abdellaoui A.
      • et al.
      Dose response of the 16p11.2 distal copy number variant on intracranial volume and basal ganglia.
      ) and 22q11.2 regions (
      • McDonald-McGinn D.M.
      • Sullivan K.E.
      • Marino B.
      • Philip N.
      • Swillen A.
      • Vorstman J.A.S.
      • et al.
      22q11.2 deletion syndrome.
      ) play a role in SCZ risk. Therefore, we consider it unlikely that the whole of SCZ cases will be explained by one commonly identifiable genetic or neurobiological signature. A more likely scenario involves a combination of those hypothesized by Kendler (2013) (
      • Kendler K.S.
      What psychiatric genetics has taught us about the nature of psychiatric illness and what is left to learn.
      ) and Liu et al., (2019) (
      • Liu X.
      • Li Y.I.
      • Pritchard J.K.
      Trans Effects on Gene Expression Can Drive Omnigenic Inheritance.
      ). For example, clusters of patients may share etiological factors falling into different scenarios of coherence, possibly with a different balance of genetic and environmental factors at the individual level. Conditions characterized by increased risk for SCZ, such as velocardiofacial syndrome (
      • Murphy K.C.
      Schizophrenia and velo-cardio-facial syndrome.
      ) and loss of function mutation in SETD1A (
      • Singh T.
      • Kurki M.I.
      • Curtis D.
      • Purcell S.M.
      • Crooks L.
      • McRae J.
      • et al.
      Rare loss-of-function variants in SETD1A are associated with schizophrenia and developmental disorders.
      ), may represent sufficient pathogenic factors, whereas other pathways may be more susceptible to gene-environment interactions. Considering that several environmental risk factors, such as bullying, childhood trauma (
      • Popovic D.
      • Schmitt A.
      • Kaurani L.
      • Senner F.
      • Papiol S.
      • Malchow B.
      • et al.
      Childhood Trauma in Schizophrenia: Current Findings and Research Perspectives.
      ) or cannabis (
      • Di Forti M.
      • Quattrone D.
      • Freeman T.P.
      • Tripoli G.
      • Gayer-Anderson C.
      • Quigley H.
      • et al.
      The contribution of cannabis use to variation in the incidence of psychotic disorder across Europe (EU-GEI): a multicentre case-control study.
      ), and their interactions and correlations with genetic risk are thought to play a role in SCZ risk (
      • Pergola G.
      • Papalino M.
      • Gelao B.
      • Sportelli L.
      • Vollerbergh W.
      • Grattagliano I.
      • et al.
      Evocative gene-environment correlation between genetic risk for schizophrenia and bullying victimization.
      ,
      • Antonucci L.A.
      • Pergola G.
      • Pigoni A.
      • Dwyer D.
      • Kambeitz-Ilankovic L.
      • Penzel N.
      • et al.
      A Pattern of Cognitive Deficits Stratified for Genetic and Environmental Risk Reliably Classifies Patients With Schizophrenia From Healthy Control Subjects.
      ), future studies with risk parsing may discover pathway-to-environment interplays much more articulated than without parsing. Additionally, the approaches we reviewed have collapsed snapshots of biological labels, regulomics, and co-expression to create models that do not integrate the dynamics of neurodevelopment. Instead, Boyce et al. (2020) elegantly presented the concept of GxExT interactions – where T stands for time (
      • Boyce W.T.
      • Levitt P.
      • Martinez F.D.
      • McEwen B.S.
      • Shonkoff J.P.
      Genes, Environments, and Time: The Biology of Adversity and Resilience.
      ). Until now data have been insufficient to examine these biological changes either in postmortem brain or other human tissues. Some encouraging results on associating pathways to treatment response based on genetic risk parsing suggest that future studies, possibly integrating time in their models, might stratify patients based on their pathway-specific risk in pharmacological trials.
      Achieving these objectives requires overcoming the current limitations of the employed approaches. Most studies focused on one or few pathways, some lacked negative controls to investigate specificity, and there may be a reporting bias favoring positive findings. The correlational nature of these associations limits the explanatory power of parsing. Only very few studies have performed biological validations in vitro of parsed risk pathways (

      Pergola G, Rampino A, Di Carlo P, Marakhovskaia A, Quarto T, Fazio L, et al. (2020): A miR-137-related biological pathway of risk for Schizophrenia is associated with human brain emotion processing.

      ,
      • Torretta S.
      • Rampino A.
      • Basso M.
      • Pergola G.
      • Di Carlo P.
      • Shin J.H.
      • et al.
      NURR1 and ERR1 Modulate the Expression of Genes of a DRD2 Coexpression Network Enriched for Schizophrenia Risk.
      ). Additionally, the prediction of gene expression in samples independent of the training cohort explains 7-11% of the actual variance in gene expression with cis-eQTL approaches (
      • Huckins L.M.
      • Dobbyn A.
      • Ruderfer D.M.
      • Hoffman G.
      • Wang W.
      • Pardiñas A.F.
      • et al.
      Gene expression imputation across multiple brain regions provides insights into schizophrenia risk.
      ) and 2-15% with co-eQTL approaches (
      • Pergola G.
      • Di Carlo P.
      • Jaffe A.E.
      • Papalino M.
      • Chen Q.
      • Hyde T.M.
      • et al.
      Prefrontal Coexpression of Schizophrenia Risk Genes Is Associated With Treatment Response in Patients.
      ,
      • Pergola G.
      • Di Carlo P.
      • D'Ambrosio E.
      • Gelao B.
      • Fazio L.
      • Papalino M.
      • et al.
      DRD2 co-expression network and a related polygenic index predict imaging, behavioral and clinical phenotypes linked to schizophrenia.
      ,
      • Fazio L.
      • Pergola G.
      • Papalino M.
      • Di Carlo P.
      • Monda A.
      • Gelao B.
      • et al.
      Transcriptomic context of DRD1 is associated with prefrontal activity and behavior during working memory.
      ). These figures suggest, (i), potential future improvement in the methods available to explain the genetic component of gene expression; (ii), that gene expression is not necessarily highly heritable for all genes across all tissues – at least not with currently available samples sizes. Furthermore, the clustering of SCZ risk genes puts together a small fraction of the putative risk genes, which in turn act on many different cell types. We expect that single nuclei sequencing will enhance the granularity of biological inferences relative to bulk tissue RNA sequencing (
      • Ruzicka B.
      • Mohammadi S.
      • Davila-Velderrain J.
      • Subburaju S.
      • Tso R.
      • Hourihan M.
      • et al.
      Single-Cell Dissection of Schizophrenia Reveals Neurodevelopmental-Synaptic Link and Transcriptional Resilience Associated Cellular State.
      ,

      Reiner BC, Crist RC, Stein LM, Weller AE, Doyle GA, Arauco-Shapiro G, et al. (2020): Single-nuclei transcriptomics of schizophrenia prefrontal cortex primarily implicates neuronal subtypes.

      ). Single cell technology developments are perhaps the most crucial step in transforming molecular data mining into actionable targets for drug development. On the phenome side, it seems likely that future studies might focus on multimodal neuroimaging data to derive a broader picture of genetic impact (
      • Antonucci L.A.
      • Fazio L.
      • Pergola G.
      • Blasi G.
      • Stolfa G.
      • Di Palo P.
      • et al.
      Joint structural-functional magnetic resonance imaging features are associated with diagnosis and real-world functioning in patients with schizophrenia.
      ).

      Conclusion

      Taken together, the investigation of parsed genetic risk combined with candidate neuroimaging intermediate phenotypes yielded significant findings with some indications of greater specificity in the genome-to-phenome mapping compared to overall genetic risk approaches. However, this is an area of active research, and methodological improvements will afford opportunities to discover mechanisms of illness, tools for subtyping, and potential novel treatments that are not available without knowledge of risk pathways. Many reports have replicated the results originated in discovery datasets, suggesting that genetic risk for SCZ has at least some coherence. These initial results represent a solid motivation to pursue further the approaches reviewed here; at the same time, increasing data availability affords opportunities to investigate time-varying and single-cell models of human brain functions.

      Uncited reference

      Rüdin E (1916): Zur Vererbung und Neuentstehung der Dementia Praecox. [On the Inheritance and Fresh Occurrence of Dementia Praecox.]. Berlin: Springer-Verlag.

      ,
      • Gottesman II,
      • Shields J.
      Schizophrenia in twins: 16 years' consecutive admissions to a psychiatric clinic.
      ,
      • Gottesman II,
      • Shields J.
      A polygenic theory of schizophrenia.
      ,
      • Gottesman II,
      • Shields J.
      Schizophrenia: geneticism and environmentalism.
      ,
      • Kety S.S.
      • Rosenthal D.
      • Wender P.H.
      • Schulsinger F.
      • Jacobsen B.
      Mental illness in the biological and adoptive families of adopted individuals who have become schizophrenic.
      ,
      • Badner J.A.
      • Gershon E.S.
      Meta-analysis of whole-genome linkage scans of bipolar disorder and schizophrenia.
      ,
      • Mah S.
      • Nelson M.R.
      • Delisi L.E.
      • Reneland R.H.
      • Markward N.
      • James M.R.
      • et al.
      Identification of the semaphorin receptor PLXNA2 as a candidate for susceptibility to schizophrenia.
      ,
      • Oldham M.C.
      • Konopka G.
      • Iwamoto K.
      • Langfelder P.
      • Kato T.
      • Horvath S.
      • et al.
      Functional organization of the transcriptome in human brain.
      ,
      • Haraldsson H.M.
      • Ettinger U.
      • Sigurdsson E.
      Developments in schizophrenia genetics: from linkage to microchips, deletions and duplications.
      ,
      • Boyle E.A.
      • Li Y.I.
      • Pritchard J.K.
      An Expanded View of Complex Traits: From Polygenic to Omnigenic.
      .

      Aknowledgements

      The authors are grateful to Dr. Daniel R. Weinberger for countless conversations on the topics of this review. This research has been supported by the project “Dopamine - dysbindin genetic interaction: a multidisciplinary approach to characterize cognitive phenotypes of schizophrenia and develop personalized treatments” (PRIN: PROGETTI DI RICERCA DI RILEVANTE INTERESSE NAZIONALE – Bando 2017 Prot. 2017K2NEF4) awarded to GP.
      Disclosures
      A.B. has received lecture fees from Otsuka, Janssen, and Lundbeck, as well as consultant fees from Biogen. All other authors have no biomedical financial interests or potential conflicts of interest to report.

      References

        • Gottesman II,
        • Shields J.
        A Polygenic Theory of Schizophrenia.
        Proceedings of the National Academy of Sciences of the United States of America. 1967; 58: 199-205
        • Gottesman II,
        • Gould T.D.
        The Endophenotype Concept in Psychiatry: Etymology and Strategic Intentions.
        American Journal of Psychiatry. 2003; 160: 636-645
        • Sullivan P.F.
        • Kendler K.S.
        • Neale M.C.
        Schizophrenia as a Complex Trait:. Evidence From a Meta-analysis of Twin Studies.
        Archives of general psychiatry. 2003; 60: 1187-1192
        • Kendler K.S.
        What psychiatric genetics has taught us about the nature of psychiatric illness and what is left to learn.
        Molecular psychiatry. 2013; 18: 1058-1066
        • Weinberger D.R.
        • Radulescu E.
        Finding the Elusive Psychiatric "Lesion" With 21st-Century Neuroanatomy: A Note of Caution.
        The American journal of psychiatry. 2016; 173: 27-33
        • Sullivan P.F.
        • Geschwind D.H.
        Defining the Genetic, Genomic, Cellular, and Diagnostic Architectures of Psychiatric Disorders.
        Cell. 2019; 177: 162-183
        • Weinberger D.R.
        Thinking About Schizophrenia in an Era of Genomic Medicine.
        The American journal of psychiatry. 2019; 176: 12-20
      1. Sullivan PF, Kendler KS, Neale MC: Schizophrenia as a Complex Trait: Evidence From a Meta-analysis of Twin Studies.

      2. Uher R, Zwicker A: Etiology in psychiatry: embracing the reality of poly‐gene‐environmental causation of mental illness.

        • Trubetskoy V.
        • Pardiñas A.F.
        • Qi T.
        • Panagiotaropoulou G.
        • Awasthi S.
        • Bigdeli T.B.
        • et al.
        Mapping genomic loci implicates genes and synaptic biology in schizophrenia.
        Nature. 2022;
        • Purcell S.M.
        • Wray N.R.
        • Stone J.L.
        • Visscher P.M.
        • O'Donovan M.C.
        • Sullivan P.F.
        • et al.
        Common polygenic variation contributes to risk of schizophrenia and bipolar disorder.
        Nature. 2009; 460: 748-752
        • Ripke S.
        • Neale B.M.
        • Corvin A.
        • Walters J.T.
        • Farh K.-H.
        • Holmans P.A.
        • et al.
        Biological insights from 108 schizophrenia-associated genetic loci.
        Nature. 2014; 511: 421-427
        • Pardiñas A.F.
        • Holmans P.
        • Pocklington A.J.
        • Escott-Price V.
        • Ripke S.
        • Carrera N.
        • et al.
        Common schizophrenia alleles are enriched in mutation-intolerant genes and in regions under strong background selection.
        Nature genetics. 2018; 50: 381-389
        • Martin A.R.
        • Daly M.J.
        • Robinson E.B.
        • Hyman S.E.
        • Neale B.M.
        Predicting Polygenic Risk of Psychiatric Disorders.
        Biological psychiatry. 2019; 86: 97-109
        • Docherty A.R.
        • Bigdeli T.B.
        • Edwards A.C.
        • Bacanu S.
        • Lee D.
        • Neale M.C.
        • et al.
        Genome-wide gene pathway analysis of psychotic illness symptom dimensions based on a new schizophrenia-specific model of the OPCRIT.
        Schizophrenia research. 2015; 164: 181-186
        • Goff D.C.
        • Tsai G.
        • Manoach D.S.
        • Coyle J.T.
        Dose-Finding Trial of D-Cycloserine Added to Neuroleptics for Negative SYmptoms in Schizophrenia.
        American Journal of Psychiatry. 1995; 152: 1213-1215
        • Goff D.C.
        • Tsai G.
        • Levitt J.
        • Amico E.
        • Manoach D.
        • Schoenfeld D.A.
        • et al.
        A Placebo-Controlled Trial of D-Cycloserine Added to Conventional Neuroleptics in Patients With Schizophrenia.
        Archives of general psychiatry. 1995; 56: 21-27
        • Guloksuz S.
        • van Os J.
        The slow death of the concept of schizophrenia and the painful birth of the psychosis spectrum.
        Psychological medicine. 2018; 48: 229-244
        • Liu X.
        • Li Y.I.
        • Pritchard J.K.
        Trans Effects on Gene Expression Can Drive Omnigenic Inheritance.
        Cell. 2019; 177: 1022-1034.e6
        • Rammos A.
        • Gonzalez L.A.N.
        • Weinberger D.R.
        • Mitchell K.J.
        • Nicodemus K.K.
        The role of polygenic risk score gene-set analysis in the context of the omnigenic model of schizophrenia.
        Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology. 2019; 44: 1562-1569
        • Hill D.P.
        • Smith B.
        • McAndrews-Hill M.S.
        • Blake J.A.
        Gene Ontology annotations: what they mean and where they come from.
        BMC bioinformatics. 2008; 9: S2
        • Rampino A.
        • Taurisano P.
        • Fanelli G.
        • Attrotto M.
        • Torretta S.
        • Antonucci L.A.
        • et al.
        A Polygenic Risk Score of glutamatergic SNPs associated with schizophrenia predicts attentional behavior and related brain activity in healthy humans.
        European neuropsychopharmacology : the journal of the European College of Neuropsychopharmacology. 2017; 27: 928-939
      3. Ripke S, Walters JTR, O’Donovan MC (2020): Mapping genomic loci prioritises genes and implicates synaptic biology in schizophrenia.

        • Maurano M.T.
        • Humbert R.
        • Rynes E.
        • Thurman R.E.
        • Haugen E.
        • Wang H.
        • et al.
        Systematic localization of common disease-associated variation in regulatory DNA.
        Science (New York, N.Y.). 2012; 337: 1190-1195
        • Potkin S.G.
        • Macciardi F.
        • Guffanti G.
        • Fallon J.H.
        • Wang Q.
        • Turner J.A.
        • et al.
        Identifying gene regulatory networks in schizophrenia.
        NeuroImage. 2010; 53: 839-847
        • Mothersill O.
        • Morris D.W.
        • Kelly S.
        • Rose E.J.
        • Fahey C.
        • O'Brien C.
        • et al.
        Effects of MIR137 on fronto-amygdala functional connectivity.
        NeuroImage. 2014; 90: 189-195
        • Hauberg M.E.
        • Roussos P.
        • Grove J.
        • Børglum A.D.
        • Mattheisen M.
        Analyzing the Role of MicroRNAs in Schizophrenia in the Context of Common Genetic Risk Variants.
        JAMA psychiatry. 2016; 73: 369-377
        • Forrest M.P.
        • Hill M.J.
        • Kavanagh D.H.
        • Tansey K.E.
        • Waite A.J.
        • Blake D.J.
        The Psychiatric Risk Gene Transcription Factor 4 (TCF4) Regulates Neurodevelopmental Pathways Associated With Schizophrenia, Autism, and Intellectual Disability.
        Schizophrenia bulletin. 2018; 44: 1100-1110
        • Ben-Shachar D.
        • Karry R.
        Sp1 expression is disrupted in schizophrenia; a possible mechanism for the abnormal expression of mitochondrial complex I genes, NDUFV1 and NDUFV2.
        PloS one. 2007; 2: e817
        • Chang H.
        • Xiao X.
        • Li M.
        The schizophrenia risk gene ZNF804A: clinical associations, biological mechanisms and neuronal functions.
        Molecular psychiatry. 2017; 22: 944-953
        • Cosgrove D.
        • Harold D.
        • Mothersill O.
        • Anney R.
        • Hill M.J.
        • Bray N.J.
        • et al.
        MiR-137-derived polygenic risk: effects on cognitive performance in patients with schizophrenia and controls.
        Translational psychiatry. 2017; 7e1012
        • Amariuta T.
        • Ishigaki K.
        • Sugishita H.
        • Ohta T.
        • Koido M.
        • Dey K.K.
        • et al.
        Improving the trans-ancestry portability of polygenic risk scores by prioritizing variants in predicted cell-type-specific regulatory elements.
        Nature genetics. 2020; 52: 1346-1354
        • Gaiteri C.
        • Ding Y.
        • French B.
        • Tseng G.C.
        • Sibille E.
        Beyond modules and hubs: the potential of gene coexpression networks for investigating molecular mechanisms of complex brain disorders.
        Genes, brain, and behavior. 2014; 13: 13-24
        • Parikshak N.N.
        • Gandal M.J.
        • Geschwind D.H.
        Systems biology and gene networks in neurodevelopmental and neurodegenerative disorders.
        Nature reviews. Genetics. 2015; 16: 441-458
        • Fromer M.
        • Roussos P.
        • Sieberts S.K.
        • Johnson J.S.
        • Kavanagh D.H.
        • Perumal T.M.
        • et al.
        Gene expression elucidates functional impact of polygenic risk for schizophrenia.
        Nature neuroscience. 2016; 19: 1442-1453
        • Gandal M.J.
        • Haney J.R.
        • Parikshak N.N.
        • Leppa V.
        • Ramaswami G.
        • Hartl C.
        • et al.
        Shared molecular neuropathology across major psychiatric disorders parallels polygenic overlap.
        Science (New York, N.Y.). 2018; 359: 693-697
        • Wang C.
        • Liu B.
        • Zhang X.
        • Cui Y.
        • Yu C.
        • Jiang T.
        Multilocus genetic profile in dopaminergic pathway modulates the striatum and working memory.
        Scientific reports. 2018; 8: 5372
        • Gallinat J.
        • Götz T.
        • Kalus P.
        • Bajbouj M.
        • Sander T.
        • Winterer G.
        Genetic Variations of the NR3A Subunit of the NMDA Receptor Modulate Prefrontal Cerebral Activity in Humans.
        Journal of Cognitive Neuroscience. 2007; 19: 59-68
        • Nuechterlein K.H.
        • Barch D.M.
        • Gold J.M.
        • Goldberg T.E.
        • Green M.F.
        • Heaton R.K.
        Identification of separable cognitive factors in schizophrenia.
        Schizophrenia research. 2004; 72: 29-39
        • Blasi G.
        • Mattay V.S.
        • Bertolino A.
        • Elvevåg B.
        • Callicott J.H.
        • Das S.
        • et al.
        Effect of catechol-O-methyltransferase val158met genotype on attentional control.
        The Journal of neuroscience : the official journal of the Society for Neuroscience. 2005; 25: 5038-5045
      4. Barbu MC, Thng G, Adams MJ, Marwick K, Grant SGN, McIntosh AM, et al. (2022): Pathway-based polygenic risk scores for schizophrenia and associations with clinical and neuroimaging phenotypes in UK Biobank.

        • Corley E.
        • Holleran L.
        • Fahey L.
        • Corvin A.
        • Morris D.W.
        • Donohoe G.
        Microglial-expressed genetic risk variants, cognitive function and brain volume in patients with schizophrenia and healthy controls.
        Translational psychiatry. 2021; 11: 490
        • Hargreaves A.
        • Anney R.
        • O'Dushlaine C.
        • Nicodemus K.K.
        • Gill M.
        • Corvin A.
        • et al.
        The one and the many: effects of the cell adhesion molecule pathway on neuropsychological function in psychosis.
        Psychological medicine. 2014; 44: 2177-2187
        • Di Biase M.A.
        • Geaghan M.P.
        • Reay W.R.
        • Seidlitz J.
        • Weickert C.S.
        • Pébay A.
        • et al.
        Cell type-specific manifestations of cortical thickness heterogeneity in schizophrenia.
        Molecular psychiatry. 2022; : 1-9
        • Akbarian S.
        • Liu C.
        • Knowles J.A.
        • Vaccarino F.M.
        • Farnham P.J.
        • Crawford G.E.
        • et al.
        The PsychENCODE project.
        Nature neuroscience. 2015; 18: 1707-1712
      5. Wang D, Liu S, Warrell J, Won H, Shi X, Navarro FCP, et al. (2018): Comprehensive functional genomic resource and integrative model for the human brain. Science (New York, N.Y.) 362.

      6. Ghosal S, Chen Q, Pergola G, Goldman AL, Ulrich W, Weinberger DR, et al. (2021): A Biologically Interpretable Graph Convolutional Network to Link Genetic Risk Pathways and Neuroimaging Markers of Disease.

      7. Chen J, Schwarz E (01/12/2017): BioMM: Biologically-informed Multi-stage Machine learning for identification of epigenetic fingerprints.

        • Chen J.
        • Zang Z.
        • Braun U.
        • Schwarz K.
        • Harneit A.
        • Kremer T.
        • et al.
        Association of a Reproducible Epigenetic Risk Profile for Schizophrenia With Brain Methylation and Function.
        JAMA psychiatry. 2020; 77: 628-636
        • Meyer-Lindenberg A.
        • Weinberger D.
        Intermediate phenotypes and genetic mechanisms of psychiatric disorders.
        nature reviews neuroscience. 2005; 7: 818-827
        • Bertolino A.
        • Blasi G.
        The genetics of schizophrenia.
        Neuroscience. 2009; 164: 288-299
        • Mandrekar J.N.
        Receiver operating characteristic curve in diagnostic test assessment.
        Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2010; 5: 1315-1316
        • Janouschek H.
        • Eickhoff C.R.
        • Mühleisen T.W.
        • Eickhoff S.B.
        • Nickl-Jockschat T.
        Using coordinate-based meta-analyses to explore structural imaging genetics.
        Brain structure & function. 2018; 223: 3045-3061
        • Gillis J.
        • Pavlidis P.
        Assessing identity, redundancy and confounds in Gene Ontology annotations over time.
        Bioinformatics (Oxford, England). 2013; 29: 476-482
        • Gillis J.
        • Pavlidis P.
        Characterizing the state of the art in the computational assignment of gene function: lessons from the first critical assessment of functional annotation (CAFA).
        BMC bioinformatics. 2013; 14: S15
        • Bleazard T.
        • Lamb J.A.
        • Griffiths-Jones S.
        Bias in microRNA functional enrichment analysis.
        Bioinformatics (Oxford, England). 2015; 31: 1592-1598
        • Bond G.L.
        • Hu W.
        • Bond E.E.
        • Robins H.
        • Lutzker S.G.
        • Arva N.C.
        • et al.
        A single nucleotide polymorphism in the MDM2 promoter attenuates the p53 tumor suppressor pathway and accelerates tumor formation in humans.
        Cell. 2004; 119: 591-602
        • Joshi M.
        • Kapopoulou A.
        • Laurent S.
        Impact of Genetic Variation in Gene Regulatory Sequences: A Population Genomics Perspective.
        Frontiers in genetics. 2021; 12660899
        • Wright C.
        • Gupta C.N.
        • Chen J.
        • Patel V.
        • Calhoun V.D.
        • Ehrlich S.
        • et al.
        Polymorphisms in MIR137HG and microRNA-137-regulated genes influence gray matter structure in schizophrenia.
        Translational psychiatry. 2016; 6: e724
      8. Pergola G, Rampino A, Di Carlo P, Marakhovskaia A, Quarto T, Fazio L, et al. (2020): A miR-137-related biological pathway of risk for Schizophrenia is associated with human brain emotion processing.

        • Cosgrove D.
        • Mothersill D.O.
        • Whitton L.
        • Harold D.
        • Kelly S.
        • Holleran L.
        • et al.
        Effects of MiR-137 genetic risk score on brain volume and cortical measures in patients with schizophrenia and controls.
        American journal of medical genetics. Part B, Neuropsychiatric genetics : the official publication of the International Society of Psychiatric Genetics. 2018; 177: 369-376
        • Yao Y.
        • Guo W.
        • Zhang S.
        • Yu H.
        • Yan H.
        • Zhang H.
        • et al.
        Cell type-specific and cross-population polygenic risk score analyses of MIR137 gene pathway in schizophrenia.
        iScience. 2021; 24102785
        • Sey N.Y.A.
        • Hu B.
        • Mah W.
        • Fauni H.
        • McAfee J.C.
        • Rajarajan P.
        • et al.
        A computational tool (H-MAGMA) for improved prediction of brain-disorder risk genes by incorporating brain chromatin interaction profiles.
        Nature neuroscience. 2020; 23: 583-593
      9. Casella AM, Colantuoni C, Ament SA (2021): Regulome-wide association study identifies enhancer properties associated with risk for schizophrenia.

        • Mahmoudi E.
        • Cairns M.J.
        MiR-137: an important player in neural development and neoplastic transformation.
        Molecular psychiatry. 2017; 22: 44-55
        • Harrison P.J.
        • Weinberger D.R.
        Schizophrenia genes, gene expression, and neuropathology: on the matter of their convergence.
        Molecular psychiatry. 2005; 10 (image 5): 40-68
        • Roussos P.
        • Katsel P.
        • Davis K.L.
        • Siever L.J.
        • Haroutunian V.
        A system-level transcriptomic analysis of schizophrenia using postmortem brain tissue samples.
        Archives of general psychiatry. 2012; 69: 1205-1213
        • Roussos P.
        • Guennewig B.
        • Kaczorowski D.C.
        • Barry G.
        • Brennand K.J.
        Activity-Dependent Changes in Gene Expression in Schizophrenia Human-Induced Pluripotent Stem Cell Neurons.
        JAMA psychiatry. 2016; 73: 1180-1188
      10. Li M, Santpere G, Imamura Kawasawa Y, Evgrafov OV, Gulden FO, Pochareddy S, et al. (2018): Integrative functional genomic analysis of human brain development and neuropsychiatric risks. Science (New York, N.Y.) 362.

        • Walker R.L.
        • Ramaswami G.
        • Hartl C.
        • Mancuso N.
        • Gandal M.J.
        • La Torre-Ubieta L de
        • et al.
        Genetic Control of Expression and Splicing in Developing Human Brain Informs Disease Mechanisms.
        Cell. 2019; 179 (e22): 750-771
        • Werling D.M.
        • Pochareddy S.
        • Choi J.
        • An J.-Y.
        • Sheppard B.
        • Peng M.
        • et al.
        Whole-Genome and RNA Sequencing Reveal Variation and Transcriptomic Coordination in the Developing Human Prefrontal Cortex.
        Cell reports. 2020; 31107489
        • Hartl C.L.
        • Ramaswami G.
        • Pembroke W.G.
        • Muller S.
        • Pintacuda G.
        • Saha A.
        • et al.
        Coexpression network architecture reveals the brain-wide and multiregional basis of disease susceptibility.
        Nature neuroscience. 2021; 24: 1313-1323
        • Jia P.
        • Manuel A.M.
        • Fernandes B.S.
        • Dai Y.
        • Zhao Z.
        Distinct effect of prenatal and postnatal brain expression across 20 brain disorders and anthropometric social traits: a systematic study of spatiotemporal modularity.
        Briefings in bioinformatics. 2021; 22
      11. Gandal MJ, Zhang P, Hadjimichael E, Walker RL, Chen C, Liu S, et al. (2018): Transcriptome-wide isoform-level dysregulation in ASD, schizophrenia, and bipolar disorder. Science (New York, N.Y.) 362.

        • Pergola G.
        • Di Carlo P.
        • Jaffe A.E.
        • Papalino M.
        • Chen Q.
        • Hyde T.M.
        • et al.
        Prefrontal Coexpression of Schizophrenia Risk Genes Is Associated With Treatment Response in Patients.
        Biological psychiatry. 2019; 86: 45-55
        • Radulescu E.
        • Jaffe A.E.
        • Straub R.E.
        • Chen Q.
        • Shin J.H.
        • Hyde T.M.
        • et al.
        Identification and prioritization of gene sets associated with schizophrenia risk by co-expression network analysis in human brain.
        Molecular psychiatry. 2020; 25: 791-804
        • Gulsuner S.
        • Walsh T.
        • Watts A.C.
        • Lee M.K.
        • Thornton A.M.
        • Casadei S.
        • et al.
        Spatial and temporal mapping of de novo mutations in schizophrenia to a fetal prefrontal cortical network.
        Cell. 2013; 154: 518-529
        • Wang Q.
        • Li M.
        • Yang Z.
        • Hu X.
        • Wu H.-M.
        • Ni P.
        • et al.
        Increased co-expression of genes harboring the damaging de novo mutations in Chinese schizophrenic patients during prenatal development.
        Scientific reports. 2015; 518209
        • Kim M.
        • Haney J.R.
        • Zhang P.
        • Hernandez L.M.
        • Wang L.-K.
        • Perez-Cano L.
        • et al.
        Brain gene co-expression networks link complement signaling with convergent synaptic pathology in schizophrenia.
        Nature neuroscience. 2021; 24: 799-809
        • Hauberg M.E.
        • Fullard J.F.
        • Zhu L.
        • Cohain A.T.
        • Giambartolomei C.
        • Misir R.
        • et al.
        Differential activity of transcribed enhancers in the prefrontal cortex of 537 cases with schizophrenia and controls.
        Molecular psychiatry. 2019; 24: 1685-1695
        • Li Z.
        • Li X.
        • Jin M.
        • Liu Y.
        • He Y.
        • Jia N.
        • et al.
        Identification of potential biomarkers and their correlation with immune infiltration cells in schizophrenia using combinative bioinformatics strategy.
        Psychiatry research. 2022; 314114658
        • Hess J.L.
        • Tylee D.S.
        • Barve R.
        • Jong S de
        • Ophoff R.A.
        • Kumarasinghe N.
        • et al.
        Transcriptome-wide mega-analyses reveal joint dysregulation of immunologic genes and transcription regulators in brain and blood in schizophrenia.
        Schizophrenia research. 2016; 176: 114-124
        • Chen J.
        • Cao H.
        • Kaufmann T.
        • Westlye L.T.
        • Tost H.
        • Meyer-Lindenberg A.
        • et al.
        Identification of Reproducible BCL11A Alterations in Schizophrenia Through Individual-Level Prediction of Coexpression.
        Schizophrenia bulletin. 2020;
        • Gamazon E.R.
        • Wheeler H.E.
        • Shah K.P.
        • Mozaffari S.V.
        • Aquino-Michaels K.
        • Carroll R.J.
        • et al.
        A gene-based association method for mapping traits using reference transcriptome data.
        Nature genetics. 2015; 47: 1091-1098
        • Hore V.
        • Viñuela A.
        • Buil A.
        • Knight J.
        • McCarthy M.I.
        • Small K.
        • et al.
        Tensor decomposition for multiple-tissue gene expression experiments.
        Nature genetics. 2016; 48: 1094-1100
        • Pergola G.
        • Di Carlo P.
        • D'Ambrosio E.
        • Gelao B.
        • Fazio L.
        • Papalino M.
        • et al.
        DRD2 co-expression network and a related polygenic index predict imaging, behavioral and clinical phenotypes linked to schizophrenia.
        Translational psychiatry. 2017; 7e1006
        • D'Ambrosio E.
        • Pergola G.
        • Pardiñas A.F.
        • Dahoun T.
        • Veronese M.
        • Sportelli L.
        • et al.
        A polygenic score indexing a DRD2-related co-expression network is associated with striatal dopamine function.
        Scientific reports. 2022; 1212610
        • Fazio L.
        • Pergola G.
        • Papalino M.
        • Di Carlo P.
        • Monda A.
        • Gelao B.
        • et al.
        Transcriptomic context of DRD1 is associated with prefrontal activity and behavior during working memory.
        Proceedings of the National Academy of Sciences of the United States of America. 2018; 115: 5582-5587
        • Braun U.
        • Harneit A.
        • Pergola G.
        • Menara T.
        • Schäfer A.
        • Betzel R.F.
        • et al.
        Brain network dynamics during working memory are modulated by dopamine and diminished in schizophrenia.
        Nature communications. 2021; 12: 3478
      12. Selvaggi P, Pergola G, Gelao B, Di Carlo P, Nettis MA, Amico G, et al. (2019): Genetic Variation of a DRD2 Co-expression Network is Associated with Changes in Prefrontal Function After D2 Receptors Stimulation. Cerebral cortex (New York, N.Y. : 1991) 29: 1162–1173.

        • Rodríguez N.
        • Gassó P.
        • Martínez-Pinteño A.
        • Segura À.-G.
        • Mezquida G.
        • Moreno-Izco L.
        • et al.
        Gene co-expression architecture in peripheral blood in a cohort of remitted first-episode schizophrenia patients.
        Schizophrenia (Heidelberg, Germany). 2022; 8: 45
        • Monaco A.
        • Monda A.
        • Amoroso N.
        • Bertolino A.
        • Blasi G.
        • Di Carlo P.
        • et al.
        A complex network approach reveals a pivotal substructure of genes linked to schizophrenia.
        PloS one. 2018; 13e0190110
        • Taurisano P.
        • Pergola G.
        • Monda A.
        • Antonucci L.A.
        • Di Carlo P.
        • Piarulli F.
        • et al.
        The interaction between cannabis use and a CB1-related polygenic co-expression index modulates dorsolateral prefrontal activity during working memory processing.
        Brain imaging and behavior. 2021; 15: 288-299
        • Antonucci L.A.
        • Di Carlo P.
        • Passiatore R.
        • Papalino M.
        • Monda A.
        • Amoroso N.
        • et al.
        Thalamic connectivity measured with fMRI is associated with a polygenic index predicting thalamo-prefrontal gene co-expression.
        Brain structure & function. 2019; 224: 1331-1344
        • Antonucci L.A.
        • Taurisano P.
        • Fazio L.
        • Gelao B.
        • Romano R.
        • Quarto T.
        • et al.
        Association of familial risk for schizophrenia with thalamic and medial prefrontal functional connectivity during attentional control.
        Schizophrenia research. 2016; 173: 23-29
        • Jaffe A.E.
        • Tao R.
        • Norris A.L.
        • Kealhofer M.
        • Nellore A.
        • Shin J.H.
        • et al.
        qSVA framework for RNA quality correction in differential expression analysis.
        Proceedings of the National Academy of Sciences of the United States of America. 2017; 114: 7130-7135
        • Parsana P.
        • Ruberman C.
        • Jaffe A.E.
        • Schatz M.C.
        • Battle A.
        • Leek J.T.
        Addressing confounding artifacts in reconstruction of gene co-expression networks.
        Genome biology. 2019; 20: 94
        • Fisher A.J.
        • Medaglia J.D.
        • Jeronimus B.F.
        Lack of group-to-individual generalizability is a threat to human subjects research.
        Proceedings of the National Academy of Sciences of the United States of America. 2018; 115: E6106-E6115
        • Alnæs D.
        • Kaufmann T.
        • van der Meer D.
        • Córdova-Palomera A.
        • Rokicki J.
        • Moberget T.
        • et al.
        Brain Heterogeneity in Schizophrenia and Its Association With Polygenic Risk.
        JAMA psychiatry. 2019; 76: 739-748
        • Bigdeli T.B.
        • Fanous A.H.
        • Li Y.
        • Rajeevan N.
        • Sayward F.
        • Genovese G.
        • et al.
        Genome-Wide Association Studies of Schizophrenia and Bipolar Disorder in a Diverse Cohort of US Veterans.
        Schizophrenia bulletin. 2021; 47: 517-529
        • Genovese G.
        • Fromer M.
        • Stahl E.A.
        • Ruderfer D.M.
        • Chambert K.
        • Landén M.
        • et al.
        Increased burden of ultra-rare protein-altering variants among 4,877 individuals with schizophrenia.
        Nature neuroscience. 2016; 19: 1433-1441
        • Singh T.
        • Poterba T.
        • Curtis D.
        • Akil H.
        • Al Eissa M.
        • Barchas J.D.
        • et al.
        Rare coding variants in ten genes confer substantial risk for schizophrenia.
        Nature. 2022;
        • Marshall C.R.
        • Howrigan D.P.
        • Merico D.
        • Thiruvahindrapuram B.
        • Wu W.
        • Greer D.S.
        • et al.
        Contribution of copy number variants to schizophrenia from a genome-wide study of 41,321 subjects.
        Nature genetics. 2017; 49: 27-35
        • Sønderby I.E.
        • Gústafsson Ó.
        • Doan N.T.
        • Hibar D.P.
        • Martin-Brevet S.
        • Abdellaoui A.
        • et al.
        Dose response of the 16p11.2 distal copy number variant on intracranial volume and basal ganglia.
        Molecular psychiatry. 2020; 25: 584-602
        • McDonald-McGinn D.M.
        • Sullivan K.E.
        • Marino B.
        • Philip N.
        • Swillen A.
        • Vorstman J.A.S.
        • et al.
        22q11.2 deletion syndrome.
        Nature reviews. Disease primers. 2015; 115071
        • Murphy K.C.
        Schizophrenia and velo-cardio-facial syndrome.
        The Lancet. 2002; 359: 426-430
        • Singh T.
        • Kurki M.I.
        • Curtis D.
        • Purcell S.M.
        • Crooks L.
        • McRae J.
        • et al.
        Rare loss-of-function variants in SETD1A are associated with schizophrenia and developmental disorders.
        Nature neuroscience. 2016; 19: 571-577
        • Popovic D.
        • Schmitt A.
        • Kaurani L.
        • Senner F.
        • Papiol S.
        • Malchow B.
        • et al.
        Childhood Trauma in Schizophrenia: Current Findings and Research Perspectives.
        Frontiers in neuroscience. 2019; 13: 274
        • Di Forti M.
        • Quattrone D.
        • Freeman T.P.
        • Tripoli G.
        • Gayer-Anderson C.
        • Quigley H.
        • et al.
        The contribution of cannabis use to variation in the incidence of psychotic disorder across Europe (EU-GEI): a multicentre case-control study.
        The Lancet Psychiatry. 2019; 6: 427-436
        • Pergola G.
        • Papalino M.
        • Gelao B.
        • Sportelli L.
        • Vollerbergh W.
        • Grattagliano I.
        • et al.
        Evocative gene-environment correlation between genetic risk for schizophrenia and bullying victimization.
        World psychiatry : official journal of the World Psychiatric Association (WPA). 2019; 18: 366-367
        • Antonucci L.A.
        • Pergola G.
        • Pigoni A.
        • Dwyer D.
        • Kambeitz-Ilankovic L.
        • Penzel N.
        • et al.
        A Pattern of Cognitive Deficits Stratified for Genetic and Environmental Risk Reliably Classifies Patients With Schizophrenia From Healthy Control Subjects.
        Biological psychiatry. 2020; 87: 697-707
        • Boyce W.T.
        • Levitt P.
        • Martinez F.D.
        • McEwen B.S.
        • Shonkoff J.P.
        Genes, Environments, and Time: The Biology of Adversity and Resilience.
        Pediatrics. 2021; 147
        • Torretta S.
        • Rampino A.
        • Basso M.
        • Pergola G.
        • Di Carlo P.
        • Shin J.H.
        • et al.
        NURR1 and ERR1 Modulate the Expression of Genes of a DRD2 Coexpression Network Enriched for Schizophrenia Risk.
        The Journal of neuroscience : the official journal of the Society for Neuroscience. 2020; 40: 932-941
        • Huckins L.M.
        • Dobbyn A.
        • Ruderfer D.M.
        • Hoffman G.
        • Wang W.
        • Pardiñas A.F.
        • et al.
        Gene expression imputation across multiple brain regions provides insights into schizophrenia risk.
        Nature genetics. 2019; 51: 659-674
        • Ruzicka B.
        • Mohammadi S.
        • Davila-Velderrain J.
        • Subburaju S.
        • Tso R.
        • Hourihan M.
        • et al.
        Single-Cell Dissection of Schizophrenia Reveals Neurodevelopmental-Synaptic Link and Transcriptional Resilience Associated Cellular State.
        Biological psychiatry. 2021; 89: S106
      13. Reiner BC, Crist RC, Stein LM, Weller AE, Doyle GA, Arauco-Shapiro G, et al. (2020): Single-nuclei transcriptomics of schizophrenia prefrontal cortex primarily implicates neuronal subtypes.

        • Antonucci L.A.
        • Fazio L.
        • Pergola G.
        • Blasi G.
        • Stolfa G.
        • Di Palo P.
        • et al.
        Joint structural-functional magnetic resonance imaging features are associated with diagnosis and real-world functioning in patients with schizophrenia.
        Schizophrenia research. 2022; 240: 193-203
      14. Rüdin E (1916): Zur Vererbung und Neuentstehung der Dementia Praecox. [On the Inheritance and Fresh Occurrence of Dementia Praecox.]. Berlin: Springer-Verlag.

        • Gottesman II,
        • Shields J.
        Schizophrenia in twins: 16 years' consecutive admissions to a psychiatric clinic.
        The British journal of psychiatry : the journal of mental science. 1966; 112: 809-818
        • Gottesman II,
        • Shields J.
        A polygenic theory of schizophrenia.
        Proceedings of the National Academy of Sciences of the United States of America. 1967; 58: 199-205
        • Gottesman II,
        • Shields J.
        Schizophrenia: geneticism and environmentalism.
        Human heredity. 1971; 21: 517-522
        • Kety S.S.
        • Rosenthal D.
        • Wender P.H.
        • Schulsinger F.
        • Jacobsen B.
        Mental illness in the biological and adoptive families of adopted individuals who have become schizophrenic.
        Behavior genetics. 1976; 6: 219-225
        • Badner J.A.
        • Gershon E.S.
        Meta-analysis of whole-genome linkage scans of bipolar disorder and schizophrenia.
        Molecular psychiatry. 2002; 7: 405-411
        • Mah S.
        • Nelson M.R.
        • Delisi L.E.
        • Reneland R.H.
        • Markward N.
        • James M.R.
        • et al.
        Identification of the semaphorin receptor PLXNA2 as a candidate for susceptibility to schizophrenia.
        Molecular psychiatry. 2006; 11: 471-478
        • Oldham M.C.
        • Konopka G.
        • Iwamoto K.
        • Langfelder P.
        • Kato T.
        • Horvath S.
        • et al.
        Functional organization of the transcriptome in human brain.
        Nature neuroscience. 2008; 11: 1271-1282
        • Haraldsson H.M.
        • Ettinger U.
        • Sigurdsson E.
        Developments in schizophrenia genetics: from linkage to microchips, deletions and duplications.
        Nordic journal of psychiatry. 2011; 65: 82-88
        • Boyle E.A.
        • Li Y.I.
        • Pritchard J.K.
        An Expanded View of Complex Traits: From Polygenic to Omnigenic.
        Cell. 2017; 169: 1177-1186