The resulting superimposed distortion map showed a significantly improved B0 field map quality (reduced spectral width and improved homogeneity).
In contrast, the reads that were derived from CNV deletions in the target genome and that were mapped to copies in the reference paralogues either had reduced mapping quality or were rejected when a unique mapping criterion was applied.
We used only the reads that were uniquely mapped to reference assemblies in the CLC and Mosaik+GigaBayes pipelines to reduce possible mapping errors, because both these read-mapping and SNP detection algorithms do not utilize mapping quality information [ 80].
A parallel analysis relying on consensus base quality, quality of adjacent bases, and read mapping quality filters yielded similar SNP calls, but the fraction of the genome compliant with the analysis criteria was slightly reduced [Additional File 1: Supplementary Table S3].
Reads with mapping quality zero were discarded.
The mapping produced a SAM formatted file containing mapping coordinates, mapping quality, and information on split reads.
All reads with mapping quality scores <20 were excluded.
Only alignments with mapping quality > 16 were kept.
This allowed SNP haplotypes occurring within the 120 base reads to align with higher mapping quality and reduce these errors.
Second, directly filtering low mapping quality reads reduces the signal-to-noise ratio (SNR) of the RC profile [ 23] so, instead, seqCNA provides a filter that discards windows, based on the mean mapping quality of proper reads.
