Change rsID to chromosome:location in bim files for plink

Question

[amitgourav.ghosh12]

Some of the SNPs in my dataset are without rs ids, so in the bim file, they have dots instead of rs ids for them. So for errorless analysis how should replace the rs id column with that of the chromosome: location column instead?

Answer 1

[zx8754]

Using awk, (not tested):

awk '{if($2 == ".") {print $1,$1"_"$3,$4,$5,$6} else {print $0}}' myfile.bim > myfileClean.bim

[Corentin]

Thanks for the answer, there is just a typo in the awk command: should be $1_$4 instead of $1_$3 (if one wants the bp coordinates in the ID, which I think would usually be the case), and there is a missing $3 as well:

awk '{if($2 == ".") {print $1,$1"_"$4,$3,$4,$5,$6} else {print $0}}' myfile.bim > myfileClean.bim

Answer 2

[chrchang523]

This can also be done with plink 2.0's --set-all-var-ids flag.

But it's more valuable to understand the awk solution, since that opens the door to a broadly useful set of text-manipulation techniques.

Answer 3

[jk587]

or use the plink option of --set-missing-var-ids @:# Note the : can be replaced by other characters