RNA Variant Calling

RNA Variant Calling

DRAGEN RNA variant calling uses the DRAGEN Somatic Small Variant Caller to call SNVs and indels. To perform variant calling, DRAGEN uses a probability model that weighs the evidence of a real variant against evidence for various noise models. If the quality score for a variant exceeds a certain threshold, then the variant is reported in the output VCF with the PASS label. DRAGEN also applies filters, such as weak_evidence and base_quality, that might indicate if the variant does not reach the thresholds required to qualify as a passing call. For more information on DRAGEN DNA somatic variant calling, see Somatic Mode for Small Variant Calling.

DRAGEN RNA also supports forced genotyping (ForceGT). In ForceGT mode, a VCF that contains variants of interest can be provided to the DRAGEN RNA Variant Caller, and the output VCF will contain all variants from the input with annotation. ForceGT might not be able to accurately call long deletions (> 50 bp) and complex variants. Complex variants are variants that require more than one substitution, insertion, or deletion event to transform the REF allele into the ALT allele.

Run RNA Variant Calling

You can use a FASTQ, BAM, or CRAM file as input. Optionally, you can provide a GTF annotation file for more accurate split junction mapping. See Rna Pipeline - Input Options for more information.

To enable RNA variant calling, set --enable-rna and --enable-variant-caller to "true". To enable ForceGT, use --vc-forcegt-vcf <forcegt_vcf_file>.

The following is an example RNA variant calling command line.

dragen \
--fastq-file1=<fastq1_file> \
--fastq-file2=<fastq2_file> \
--RGID=<read_group_id> \
--RGSM=<read_group_sample_name> \
--enable-duplicate-marking=true \
--dupmark-version=hash \
--enable-rna=true \
--enable-variant-caller=true \
--ref-dir=<ref_hashtable_dir> \
--output-directory=<output_dir> \
--output-file-prefix=<output_prefix> \
--annotation-file=<gtf_annotation_file> \
--vc-forcegt-vcf=<forcegt_vcf_file> 

Output files

RNA variant calling outputs a VCF file that includes PASS variants and variants that did not pass, due to filters or weak evidence. For more information on filters and additional command line options, see Somatic Mode for Small Variant Calling.

DRAGEN RNA applies a heuristic to genotype the variants based on the variant allele frequency: if the AF is at least 85%, then the GT field will be set to 1/1. Otherwise GT will always be reported as 0/1. This behavior and threshold can be adjusted with the following options:

• --rna-vc-enable-homozygous-genotype (default=true)

• --rna-vc-homozygous-genotype-af-threshold (default=0.85)