Fig. 4

Frequency distribution of the base quality from each method. Blue line: the average frequency distribution of true minor alleles with the 1000 weakest signals; red line: the average frequency distribution of sequencing errors with the 1000 strongest signals (see main text for the explanation of weakest and strongest signals). Note that with GATK the weakest true minor alleles in general have lower recalibrated base quality scores than the strongest sequencing errors, while with SEGREG the distributions of recalibrated base quality scores are more similar for the weakest true minor alleles vs. the strongest sequencing errors