Skip to main content


Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

Fig. 2 | BMC Genomics

Fig. 2

From: A machine learning model to determine the accuracy of variant calls in capture-based next generation sequencing

Fig. 2

Key model parameters. Predictive power of call-dependent and site-dependent signals. All instances of the data set are shown; instances marked in green are NGS calls that were confirmed with Sanger, and instances marked in red are cases that did not confirm. a Call quality (QUAL), a single metric of call quality as measured by GATK HaplotypeCaller. Left panel: all variants. Right panel: zoom-in of variants which did not confirm, showing that some not confirmed variants have relatively high QUAL scores. b The values of the two strongest features associated with the call: allele frequency (AF) and QUAL. c The values of QUAL normalized by read depth (QD), AF, and weighted homopolymer rate (WHR)

Back to article page