Example of using GEMINI genotype columns to identify confident somatic mutations in a cancer experiment

gemini-somatic.sh

# DOCS:
# https://gemini.readthedocs.org/en/latest/content/database_schema.html#genotype-information

# GEMINI SOURCE:
# https://github.com/arq5x/gemini

#########################################################################
# load a VCF for a tumor / normal pair into gemini.
# - use 4 cores
# - assume VCF has been annotated with snpEff
#########################################################################
$ gemini load -v tumor-normal.vcf -t snpEff --cores 4 tumor-normal.vcf.db


#########################################################################
# Identify novel somatic mutations in the tumor that are likely to
# impact gene function.
#
# SOMATIC:
#   1. the tumor has an alternate allele (HET)
#   2. the normal does not (HOM_REF)
#   3. Also, the normal has NO evidence for the alternate allele.  To
#      enfore this, we require that the were 0 alignments for the normal
#      with the alternate allele.
#########################################################################

$ gemini query -q "select chrom, start, end, ref, alt, type, gene \
                   from variants 
                   where impact_severity !='LOW' 
                   and in_dbsnp = 0" \
               --gt-filter "gt_types.TUMOR == HET and 
                            gt_types.NORMAL == HOM_REF and 
                            gt_alt_depths.NORMAL == 0" \
               tumor-normal.vcf.db