removing duplicate SNPs (same position) with lowest call rate

Question

[jani.p.heikkinen]

I am trying to solve a problem with my genotyped array data set. For reason or another, the data set has duplicate or with three different names pointing to the same position. For example:

index  SNP    pos        A1  A2  F_MISS
2046   snp_1  113890304  C   T   0
2047   snp_2  113890304  C   T   0.000422
2048   snp_3  113890304  C   T   0

I want to build a list for SNP names to be removed (so I can exclude them in PLINK).

So from the SNPs above, snp_1 or snp_3 and snp_2 should be in removal list.

How would I achieve this?

Answer 1

[TriS]

If you just want to remove duplicates in R (not tested):

name_position <- apply(mySNPmatrix,1,function(x) paste(x[2],x[3],sep="_"))
mySNPmatrix <- mySNPmatrix[-which(duplicated(name_position)),]

However, it seems that the F_MISS col is not duplicated, so pay attention to that when removing rows

Answer 2

[christopher medway]

Try this bash one-liner (not tested). You may need to lose the header line though.

sort -k 3 -k 6 input.txt | awk '!seen[$3]++' | awk '{print $2}' > output.txt