SNP Calling

mpileup (SAMtools)

• Create bcf file (binary version of vcf) from alignment file (bam)

  #Use Q to set base quality threshold
  samtools mpileup -Q 25 -ugf <reference.fasta> <file.bam> | bcftools view -bvcg - > accepted_hits.raw.bcf
  #mpileup BAQ computation might be too stringent, disabling BAQ may be needed
  samtools mpileup -B -Q 25 -ugf <reference.fasta> <file.bam> | bcftools view -bvcg - > accepted_hits.raw.bcf

• Convert bcf to vcf (vairant call format) and filter, if needed

#filter using varFilter (from vcfutils)
  #use Q to set mapping quality
  #use d for minimum read depth
  bcftools view accepted_hits.raw.bcf | vcfutils.pl varFilter -d 5 -Q 20 >| accepted_hits.raw.vcf

Filtering SNPs

• Remove, or retrieve, known SNPs

  #To remove known SNPs, use intersectBed (bedtools).  Known SNPs can be downloaded from Ensembl or NCBI/UCSC (eg. dbSNP)
  intersectBed -a accepted_hits.raw.filtered.vcf -b Mus_musculus.NCBIM37.60.bed -wo > filteredSNPs.vcf

Annotating SNPs

• Determining the effect of the SNP

snpEff

 #Use appropriate i) organism  and ii) annotation (eg. UCSC, RefSeq, Ensembl)
  #Mouse: mm37 (UCSC/RefSeq), mm37.61 (Ensembl)
  #Human: hg37 (UCSC/RefSeq), hg37.61 (Ensembl)
  #Output is created in several files: an html summary file and text files with detailed information 
  snpEff -vcf4 mm37.61 accepted_hits.raw.filtered.vcf > accepted_hits.raw.filtered.snpEff

Resources from the Broad Institute

• ​Best Practices For Variant Calling With The GATK (workshop materials)