seXY

seXY infers biological sex from genotype array data to provide quality control for genome-wide association studies (GWAS).

Key Features:

• Input data: Operates on genotype array data for sample-level sex inference.
• Model: Implements a logistic regression model to classify sex.
• Genomic indicators: Integrates X chromosome heterozygosity and Y chromosome missingness as model predictors.
• Validation: Demonstrated accuracy exceeding 99.5% in cross-validation tests on cohorts of 889 males and 5,361 females from prostate cancer and ovarian cancer GWAS datasets.
• Comparative performance: Shows marginally better male classification and a 3% improvement in female classification accuracy relative to PLINK's X chromosome heterozygosity–based method.

Scientific Applications:

• GWAS quality control: Detects discordance between reported and inferred sex to improve sample QC in genome-wide association studies.
• Sex concordance checks: Verifies reported sex against genotype-derived sex to reduce misclassification bias in genetic analyses.
• Study-specific validation: Validated on prostate cancer and ovarian cancer GWAS datasets for large-cohort sex inference assessment.

Methodology:

Uses a logistic regression model that combines X chromosome heterozygosity and Y chromosome missingness, with performance assessed by cross-validation on the specified GWAS cohorts.