BAMClipper: removing primers from alignments to minimize false-negative mutations in amplicon next-generation sequencing

Abstract

Abstract Amplicon-based next-generation sequencing (NGS) has been widely adopted for genetic variation detection in human and other organisms. Conventional data analysis paradigm includes primer trimming before read mapping. Here we introduce BAMClipper that removes primer sequences after mapping original sequencing reads by soft-clipping SAM/BAM alignments. Mutation detection accuracy was affected by the choice of primer handling approach based on real NGS datasets of 7 human peripheral blood or breast cancer tissue samples with known BRCA1/BRCA2 mutations and >130000 simulated NGS datasets with unique mutations. BAMClipper approach detected a BRCA1 deletion (c.1620_1636del) that was otherwise missed due to edge effect. Simulation showed high false-negative rate when primers were perfectly trimmed as in conventional practice. Among the other 6 samples, variant allele frequencies of 5 BRCA1/BRCA2 mutations (indel or single-nucleotide variants) were diluted by apparently wild-type primer sequences from an overlapping amplicon (17 to 82% under-estimation). BAMClipper was robust in both situations and all 7 mutations were detected. When compared with Cutadapt, BAMClipper was faster and maintained equally high primer removal effectiveness. BAMClipper is implemented in Perl and is available under an open source MIT license at https://github.com/tommyau/bamclipper.