Reads used for alignment

The alignment process involves aligning reads produced by the pipeline to a reference sequence and extracting metrics from those alignments. By default, Torrent Suite™ Software aligns all reads to the reference; however, there may be situations, particularly with large reference sequence, when the alignment exceeds the amount of time that you are willing to wait. If so, the software can define on a per-reference basis the maximum number of reads that are aligned from a run.

When the number of reads in a run exceeds a reference-specific maximum, a random sample of reads is taken and results are extrapolated to represent the full run. By sampling a random pool of reads and extrapolating the alignment quality values to represent all reads, the software provides sufficient information to assess the quality of the sample, library, and sequencing run.

The output of the alignment process is a BAM file. The BAM file includes an alignment of all reads, including the unmapped reads, with exactly one mapping per read. When a read maps to multiple locations, the mapping with the best mapping score is used. If more than one such mapping exists, a random mapping is used and a mapping quality of zero is given.