For Ion Proton dataset, loci with at least 5 reads supporting alternative base are included in the evaluation. For Illumina MiSeq dataset, filter 2 used by UDT-Seq was applied which requires > = 0.2 % frequency for alternative bases. However, the other filters were not used. We relied on the PSEM framework to properly address sequencing problems, for example, uneven depth and local sequence context induced errors.
