Extracting a panel of SNPs from a whole genome sequence

[vl250]

I'm looking to use whole genomes we already have available as part of a GWAS with genotyping data from a SNP array.

How best can I extract the relevant SNPs from my whole genomes? I have a .bed file set up with all my SNP locations, and have been Googling like mad trying to find ways to do this. Am I best to extract these from my aligned .bam file using something like:

$ bedmap --echo --fraction-map 1 <(bam2bed < reads.bam) intervals.bed > answer.bed

or

$ samtools mpileup -ugf ref.fa -l intervals.bed sample1.bam | bcftools call -vmO z -o answer.vcf.gz

Or to just restrict the final combined vcf for the whole genomes with something like this:

$ java -jar GenomeAnalysisTK.jar -T SelectVariants -R ref.fa -V all_samples.vcf -L intervals.bed -sn Sample1 -sn Sample2 -sn Sample3

I'm not sure what the advantages/disadvantages of doing it each stage would be, and how best I can get as close to a SNP array output as possible (ie with the variant present at each location).

Googling usually answers all my questions but not today!

[Gopo]

I would call SNPs with your desired variant caller (I prefer CallVariants from BBTools/BBMap) then use bedtools intersect (-a study.vcf.gz -b bed-file-of-snps) and the output is a VCF file of intersecting SNPs. Note that you might have some indels after bedtools intersect that you need to filter out with bcftools or vcftools.

If you want to use mpileup, samtools mpileup is deprecated so use bcftools mpileup.

bcftools mpileup -Ou -f <ref.fa> <sample1.bam> <sample2.bam> <sample3.bam> | bcftools call -vmO z -o <study.vcf.gz>

above example command from http://www.htslib.org/workflow/