Approach for detecting loci with systematic sequence bias: Alternative allele fractions are collected from a cohort of individuals at every locus and aggregated into allele-specific summary statistics. These aggregate allelic fractions and standard deviations at each genomic position are stored in the Incremental Database (IncDB), which does not contain any patient-specific information. The 99.9% confidence interval for expected standard deviation at each allelic fraction was generated. Genomic positions where the observed standard deviation was below the confidence interval expected were cataloged as “suspect loci” and mapped to variant calls in clinically relevant genes. Prioritization of genes for diagnostic and reporting purposes can be adjusted according to the presence of suspect loci.
