Genes Associated with Mental Illness
The genetic architecture of major mental illnesses involves both shared pleiotropic genes across multiple disorders and disorder-specific variants, with the 22q11.2 deletion representing the strongest single genetic risk factor for schizophrenia (20-fold increased risk), while common variants in genes like COMT, SLC6A4 (5-HTTLPR), MAOA, BDNF, and LINC00461 show pleiotropic effects across schizophrenia, bipolar disorder, major depression, ADHD, and autism spectrum disorder.
Key Genetic Findings by Disorder
22q11.2 Deletion Syndrome - Strongest Single Risk Factor
The 22q11.2 deletion confers a greater than 20-fold increased risk for schizophrenia compared to the general population, with approximately 1 in 4-5 adults with this deletion developing schizophrenia 1, 2. This represents the most clinically significant single genetic risk factor identified for any major psychiatric disorder. The deletion also increases risk for:
- Anxiety disorders (2-3 times population prevalence)
- Autism spectrum disorders and ADHD (persisting from childhood)
- Intellectual disability (30-40% have mild ID with IQ 55-69) 2
Notably, bipolar disorder and major depression occur at similar rates to the general population in 22q11.2DS 2.
Shared Pleiotropic Genes Across Multiple Disorders
Recent large-scale genomic analyses have identified 238 pleiotropic loci affecting multiple psychiatric disorders 3. The most significant shared genetic variants include:
LINC00461 (long non-coding RNA on chromosome 5q14.3): Shows the highest pleiotropic effects across five psychiatric traits, with functional studies supporting its vital role in neurodevelopment 4.
Serotonin System Genes:
- SLC6A4 (5-HTTLPR): The serotonin transporter gene polymorphism is implicated across major depression, bipolar disorder, and schizophrenia. The short allele variant shows gene-environment interactions with stress exposure 5, 6, 7. However, findings are inconsistent regarding which allele (short vs. long) confers risk, with socioeconomic status appearing to moderate effects 5.
Dopamine and Monoamine System Genes:
- COMT (Val158Met): The low-activity Met/Met genotype is associated with increased risk for postpartum depression and shows pleiotropic effects across multiple disorders 5, 7. This gene is involved in dopamine and noradrenaline metabolism.
- MAOA (monoamine oxidase-A): Low-activity variants increase depressive symptoms, particularly when combined with low-activity COMT variants 5, 7.
- DRD4 and DAT1: Dopamine receptor and transporter genes show associations across ADHD and other disorders 7.
Synaptic and Neuronal Function Genes:
- BDNF (Val66Met): Brain-derived neurotrophic factor polymorphism shows shared effects across multiple disorders 7.
- MAPT: Associated with schizophrenia and Parkinson's disease in astrocytes 8.
- FLOT1: Implicated in excitatory and inhibitory neurons for schizophrenia, bipolar disorder, and major depression 8.
- SCN2A and UNC5D: Involved in nervous system development and shared between ADHD and schizophrenia 9.
Disorder-Specific Patterns
Schizophrenia and Bipolar Disorder (SB Factor): These disorders share the most genetic overlap, with the majority of genetic signal being common between them 3. The shared genetic signal is substantially enriched in genes expressed in excitatory neurons 3. Only 0.4% of schizophrenia-associated SNPs and 2.3% of bipolar disorder SNPs colocalize with ADHD variants, indicating substantial genetic distinction from neurodevelopmental disorders 9.
Internalizing Disorders (Major Depression, PTSD, Anxiety): These disorders form a distinct genetic factor with high polygenic overlap and are associated with oligodendrocyte biology 3. The estrogen receptor gene (ESR1) shows specific associations with postpartum depression, particularly in interaction with 5-HTTLPR 5.
Autism Spectrum Disorder:
- PTEN mutations: Germline heterozygous PTEN mutations are associated with ASD, particularly in cases with macrocephaly and intellectual disability 10.
- Shows less genetic overlap with other psychiatric disorders compared to the high correlation seen between schizophrenia and bipolar disorder 2.
ADHD: Shows genetic sharing with schizophrenia and bipolar disorder through specific variants (3 of 12 ADHD SNPs colocalize with schizophrenia, 1 with bipolar disorder), but maintains substantial genetic distinctiveness 9.
Cellular and Functional Mechanisms
Cell-Type Specificity: Brain-derived risk genes exhibit significantly higher cell-type specificity than blood-derived genes, emphasizing the focused cellular context of genetic effects in the central nervous system 8. Key findings include:
- Microglia: APTX gene implicated in schizophrenia
- Astrocytes: MAPT for schizophrenia and Parkinson's disease
- Excitatory and inhibitory neurons: FLOT1 for schizophrenia, bipolar disorder, and major depression 8
Biological Pathways: Shared genetic signal across all 14 psychiatric disorders is enriched for:
- Transcriptional regulation (broad processes)
- Synaptic function
- Immune regulation 3
More specific pathways operate at the level of individual disorder factors, with neuronal dysfunction and immune system dysregulation representing convergent mechanisms 8.
Critical Clinical Caveats
Predictive Limitations: Despite statistically significant associations, individual genetic variants have extremely limited clinical predictive value. For example, the strongest schizophrenia risk variant (rs9268895) only raises individual risk from 1 in 196 to 1 in 167—far from clinically actionable 11. Even with 99% sensitivity and specificity, screening for low-prevalence conditions like schizophrenia yields false positive rates exceeding 50% 11.
Gene-Environment Interactions: Many genetic effects are moderated by environmental factors:
- 5-HTTLPR effects depend on stress exposure, socioeconomic status, and timing of assessment 5
- Cannabis use conveys additional risk for psychotic functioning in genetically vulnerable individuals 1
- Early marijuana use should be avoided in high-risk populations (e.g., 22q11.2DS) 2
Methodological Considerations: The genetic architecture differs substantially between ancestral populations, and most studies have been conducted in European ancestry populations, limiting generalizability 12. Machine learning approaches for genetic prediction show highly variable performance (0.48-0.95 AUC) with widespread risk of bias in study designs 13.