QIAGEN powered by

Some larger single deletions reported as multiple, individual deletions when affine gap scoring used

Issue description

When using affine gap scoring in Map Reads to References, Map Reads to Contigs, RNA-Seq Analysis, or Add Reads to Contig on affected software versions, sub-optimal alignments could occur, where opening a gap was not being sufficiently penalized and gap extension should have been preferred. Variants are still called in the right position(s) when variant detection is done using such mappings as input, but in affected regions, multiple, individual deletions may be reported where a single, larger deletion would have been more appropriate.

This problem is expected to predominantly affect mappings with lower quality, long reads, for example, PacBio or Oxford Nanopore data, or when using reference sequences from species different to the one the reads originated from.

By default Map Reads to References, Map Reads to Contigs, RNA-Seq Analysis and Add Reads to Contig use linear gap scoring, which is not affected by this problem. However, affine gap scoring is the default for these tools when included in Ready-to-Use workflows delivered by the Biomedical Genomics Workbench and by the plugins QIAseq Targeted Panel Analysis and QIAGEN GeneRead Panel Analysis. Affine gap scoring is also the default in at least some of the workflows delivered with the CLC Microbial Genomics Module.

Affected software versions

This problem will also affect the CLC Genome Finishing Module installed on the above software versions, as it delivers the tool Add Reads to Contig. In addition, any workflows delivered by plugins installed on these versions that include Map Reads to References, Map Reads to Contigs or RNA-Seq Analysis and use affine gap scoring will be affected.

This issue was fixed in CLC Genomics Workbench 12.0, CLC Genomics Server 11.0. Workflows delivered by any version of the Biomedical Genomics Analysis plugin or by  CLC Microbial Genomics Module 4.0 and higher are therefore not affected by this problem.