Rachel Soemedi, PhD

Sex differences in human health and disease are common, but severely understudied. This gap of knowledge prevents accurate risk assessment and stratification for effective and safe therapeutic interventions for sex-biased diseases. Historically, females are more often disadvantaged from male-predominant clinical study participants. Recent re-analyses of previous genome-wide association studies (GWAS) uncovered many sex-specific and sex-asymmetric disease-associated loci that are widespread throughout the genome. This suggests that genetic landscapes that shape complex disease traits are sex divergent. Recent evidence also indicates thousands of genes being sex-differentially expressed, both quantitatively (expression level) and qualitatively (splicing patterns), and they exhibit significant tissue specificity. We hypothesize that sex divergent genomes are inherent in human populations and manifest as sex-asymmetric variant allele frequency and sex-biased disease-associated haplotypes. We investigate the characteristics of the sex-biased variants (SBV) in either allele frequency (freq-SBVs) and/or phenotypic association (phe-SBVs) in gnomAD, UK Biobank and Japan Biobank datasets. We found that freq-SBVs are highly shared between different populations and they are enriched in known tissue-specific sex-differentially expressed genes. The direction and magnitude of sex biased allele frequency of shared freq-SBVs are also significantly correlated between populations. The highly shared freq-SBVs are predominantly found in transcription factor binding sites (TFBS) that are known to regulate genes involved in embryonic developments, including heart morphogenesis, central nervous system and skeletal muscle tissue developments. They have primary biological roles in alternative splicing (different isoforms from a single gene) and major pathways involved in drug metabolisms, cardiovascular disorders, cancer, autoimmune and psychiatric diseases. Interestingly, ~10% of known disease variants that are present in gnomAD are freq-SBVs. We explore the various measures of freq-SBVs and phe-SBVs, and plan to validate freq/phe-SBVs in the transcriptomic levels at 50 different human tissues, by utilizing the allele-specific expression data from the GTEx consortium.