convert 9million Marker Names to rsid

Question

[mk19726]

I have been sent GWAS summary data that I want to use for Mendelian Randomization studies. However it only contains marker name (chr:pos) rather than SNP rsid. There are 9million observations.

Is there an easy/quick way to convert them to rsid?

I am a first year PHD student with a non-programming background, so fairly simple explanations greatly appreciated!

Thank you very much for your help, Daniel

[Emily]

Does it also have alleles?

[mk19726]

yes, it does have effect and other allele

Answer 1

[Pierre Lindenbaum]

extract CHROM:POS from dbsnp, convert to CHROM:POS,RSID , sort on CHROM:POS (time consumming)...

wget -O - "ftp://ftp.ncbi.nih.gov/snp/organisms/human_9606/VCF/00-All.vcf.gz" | gunzip -c | awk -F '\t' '/[^#]/ {printf("%s:%s,%s\n",$1,$2,$3)}' | LC_ALL=C sort -T . -t, -k1,1 > dbsnp.csv

sort your data on CHROM:POS. Assuming the chrom notation is the same as dbsnp ('1' not 'chr1'), assuming the genome build is the same as dbsnp.

LC_ALL=C sort -T . -t, -k1,1 yourlist.txt > sorted.csv

join both list

LC_ALL=C join -t, -1 1 -2 1 dbsnp.csv sorted.csv > output.txt

Answer 2

[Emily]

Easiest thing is to stick the variants into the Ensembl VEP. It will calculate the consequences on genes and give you rsIDs of known variants at the loci.