Luis Barrera
Harvard Medical School
Disease-associated mutations and coding variants in human transcription factors
Gene expression in humans is modulated by thousands of genetic variants, many of which have been identified by eQTL mapping. However, eQTL methods have limited power for finding trans regulatory variants, particularly if they are rare. Here, we use a complementary approach to identify and characterize genetic variants that alter key residues in the DNA-binding domains of transcription factors. We developed a computational pipeline that uses co-crystal structures and literature annotations to prioritize non-synonymous SNPs in human populations that are likely to have an effect on DNA binding. We performed over 100 protein-binding microarray experiments to compare the DNA-binding specificity and affinity of the prioritized mutant alleles with the reference alleles. In parallel, we measured changes in DNA binding for a set of transcription factor alleles with known Mendelian mutations. While nsSNPs typically led to more conservative changes in DNA binding, we observed a continuum of effects, particularly in binding affinity changes. Our approach identified novel SNPs with potential effects on gene expression and provides insight into the molecular mechanisms of disease-causing mutations.
