Methods for detecting variants in next-generation sequencing genomic data

Abstract

A genomic data analyzer may be configured to detect and characterize, with a variant calling module, genomic variant scenarios in sequencing reads from an enriched patient genomic sample comprising a combination of a first repeat pattern and a second repeat pattern, such as repeats of homopolymer (single nucleotide) and/or heteropolymer (multiple nucleotide) basic motifs. The variant calling module may estimate the probability distribution of the length of the first repeat pattern and the probability distribution of the length of the second repeat pattern by comparing the distribution of the repeat pattern length measurements in patient data to the distribution of the repeat pattern length measurements in control data, in order to remove biases possibly induced by the next generation sequencing laboratory setup both in control and patient data. The variant calling module may further measure, read by read, the joint probability distribution for the first and the second repeat patterns lengths, and compare it with the expected joint probability distribution for various genomic variant scenarios for the patient, each variant scenario being characterized by a first length of the first repeat pattern and a second length of the second repeat pattern, to select the most likely patient genomic variant scenario as the scenario for which the measured joint probability distribution best matches the expected joint probability distribution.