Genome-wide polygenic scoring for a 14-year long-term average depression phenotype

Area of Science:

• Psychiatric Genetics
• Genomics and Epidemiology
• Depression Research

Background:

• Moderate heritability of depression contrasts with limited identification of robust genetic predictors.
• Previous studies may be limited by single-time-point phenotypic measures and small-effect genetic loci.

Purpose of the Study:

• To assess the predictive power of polygenic scores (PS) for depression using a long-term average phenotype.
• To investigate if integrating 14 years of depression data improves genetic prediction accuracy.

Main Methods:

• Developed a 14-year average depression measure from the Nurses' Health Study (N=6989).
• Estimated polygenic scores (PS) using internal and external weighting algorithms (NHS-GWAS-PS, GAIN-MDD-PS, PGC-MDD-PS).
• Assessed PS association with the long-term depression phenotype via linear, logistic, and quantile regressions.

Main Results:

• Three PS approaches explained a maximum of 0.2% of the variance in the long-term depression phenotype.
• Quantile regressions revealed larger PS impacts at higher depression symptom quantiles (75th vs. 25th percentile).
• Significant differences in PS effects were observed at the 25th and 75th percentiles for PGC-MDD-PS (P=0.03) and borderline for GAIN-MDD-PS (P=0.05).

Conclusions:

• Integrating 14 years of phenotypic data and employing various PS approaches did not substantially enhance depression genetic prediction.
• Polygenic score effects appear more significant in individuals with higher depression symptom severity.
• Unobserved vulnerability factors likely contribute to depression risk in individuals with high symptom scores.