Identifying Genetically Distinguishable Subgroups within Major Depressive Disorder

Major depressive disorder (MDD) has long been theorized to reflect an overly broad disorder class that collapses across heterogenous risk pathways. A rate limiting factor to examining the divergent validity of MDD subtypes using genomic methods is a lack of sufficiently powered data. As part of the proposed project, we will utilize Co-I Dr. Lewis’ role as a leader in depression genetics and co-chair of the PGC MDD working group to put together the largest genome-wide association study (GWAS) yet performed for various MDD subtypes, including sex-stratified, atypical, postpartum, and severe MDD. In addition, we will employ Genomic LOSEM, a novel method introduced in the grant for examining non-linear changes in genetic signal that we will use to examine how different ages at onset and socioeconomic status shift MDD genetic architecture. The subtype GWAS and Genomic LOSEM package will be made available as public resources. Standard univariate approaches that focus strictly on either meta-analyzing across MDD in all its forms or analysis of a particular subtype are unable to parse genetic risk pathways that are broadly relevant to MDD from those that are unique to a specific subtype. In addition, family-based approaches are pragmatically limited to examining a handful of subtypes at a time and cannot describe underlying biology. Genomic Structural Equation Modeling (Genomic SEM) is an innovative, multivariate framework developed by the grant PI Dr. Grotzinger for modeling genetic overlap derived from GWAS data. The well-powered GWAS of MDD subtypes will be used as input to Genomic SEM models that will formally disambiguate shared and subtype-specific genetic signal. A unique advantage of Genomic SEM is that even mutually exclusive subtypes can be included in the same statistical model. The remaining analyses will characterize subtype-specific genetic signal at varying levels of biological granularity, including estimating genetic overlap with clinically relevant external correlates (e.g., cognition, other psychiatric disorders). By applying Stratified Genomic SEM, a novel extension for estimating multivariate functional enrichment, we will characterize biological pathways involved in subtype specific risk. These biological pathways can include, for example, genes expressed early in development, in certain brain regions, or in specific types of neurons. At the gene expression level, Transcriptome-wide SEM will be used to identify the lists of genes uniquely associated with an MDD subtypes. These results will be cross-referenced with the Connectivity Map drug repurposing dataset to identify existing pharmacological interventions that may have therapeutic benefit. By utilizing sex-stratified GWAS summary statistics we will explicitly consider biological sex as a moderator of relevant genetic pathways. In addition, expanding African, East Asian, and LatinX ancestry GWAS datasets, LD-scores, functional annotations, gene expression weights and cross-ancestry methods will allow us to extend the grant aims across diverse samples. Our analyses will collectively provide the most comprehensive evaluation to-date of subtype-specific etiology within MDD. Project Number: 1R01MH141707-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of Mental Health (NIMH) | Principal Investigator: Andrew Grotzinger | Institution: UNIVERSITY OF COLORADO, Boulder, CO | Award Amount: $730,338 | Activity Code: R01 | Study Section: Neurological, Mental and Behavioral Health Study Section[NMBH] View on NIH RePORTER: https://reporter.nih.gov/project-details/11219386