We use the Broad GATK best practices for RNAseq variant calling detailed here:
https://gatk.broadinstitute.org/hc/en-us/articles/360035531192?id=3891
You can see the actual code we use here:
https://github.com/soccin/BIC-variants_pipeline
The entry point for RNAseq Variants Calling is here:
variants_pipeline_rnaseq_wrapper.pl
The key steps (besides using STAR to align in stead of BWA is)
Using
SplitNCigarReads
To deal with reads that span introns. This creates a BAM that the HaplotypeCaller can handle. The two extra options we give to Haplotype are
-dontUseSoftClippedBases
-stand_call_conf 20.0
We do Hard Filtering as recommend here.
