id summary reporter owner description type status priority milestone component resolution keywords cc 1199 Implement Variant calling pipeline Nicklas Nordborg Nicklas Nordborg "**Updated description**\\ Variant calling is started from Hisat alignment. The first step is a ""raw"" variant calling using mosdepth and !VarDict. The variant calling only depends on the same reference genome that was used for alignment. We save two files (`variants-callable.bed` and `variants-raw.vcg.gz`) from this step and attaches them to the same Hisat alignment. The seconds step is an annotation step and a filtering step. Multiple programs are used to extract and merge data from a lot of external sources (for example dbSnp, Cosmic, Swegene, etc.) This step creates a child rawbioassay item with two more files. An annotated VCF (`variants-annotated.vcf.gz`) that is the same as the raw VCF with added information from the other database, and a filtered VCF (`variants-filtered.vcf`) with the variants that survive filtering. The variants calling wizard can be executed in two modes: raw-only mode will do only the first step and the full mode will do both steps. In full mode, the raw variant calling can be skipped if it has already been done before. **Original description**\\ The starting point can probably be Hisat alignments. It is not yet sure what kind of data that is going to be produced. We will at least get a vcf file containing found variants. It might be possible to attach this data to the current alignment item, but it may also be better to create a child item for this. A child item is better if we get a lot of data files and is also required if we want to re-run the analysis without overwriting the first result. We can start building the framework for hooking into auto-confirmation and the manual wizards. It can probably be very similar to how the framework for the mBAF pipeline is implemented (#1068)." task closed major Reggie v4.24 net.sf.basedb.reggie fixed