Figure 3

Pre and post filtering variants data. Plot of the coverage (times a single sequence is read by the sequencing equipment) and the frequency of an observed variation. In reality, genomic variation occurs at a frequency of either 50% or 100% of total reads. The top figure is the distribution when no filters are applied to discriminate between sequencing errors and real variants; the bottom figure is the distribution where several filters (frequency filter, Q score filter, coverage filter and homopolymer filter) are applied to discriminate the real variations from the sequencing errors. Vertical grey line indicates 40 × coverage; horizontal grey lines indicate respectively 33%, 67% and 95% variation frequency.