This is a test version of Biostars. For the public version, visit https://www.biostars.org.
ERROR: Stopping due to mis-matching SNPs -- check +/- strand?

Hello,

I'm trying to perform HapMap PCA, after which plink bails and gives me the following message:

ERROR: Stopping due to mis-matching SNPs -- check +/- strand?

In my script I do the following to try and rectify this.

# initial merge study and hapmap data                                                                                           
plink --bfile ${DATA} --bmerge ${HAPMAP_DATA}.bed ${HAPMAP_DATA}.bim ${HAPMAP_DATA}.fam --make-bed --out clean_data/hapmap_pca/dataset/${NAME}_HapMap3_pruned

plink --bfile ${DATA} --flip clean_data/hapmap_pca/dataset/${NAME}_HapMap3_pruned.missnp --make-bed --out clean_data/hapmap_pca/dataset/${NAME}_flip

# merge study-flip and hapmap data                                                                                              
plink --bfile clean_data/hapmap_pca/dataset/${NAME}_flip --bmerge ${HAPMAP_DATA}.bed ${HAPMAP_DATA}.bim ${HAPMAP_DATA}.fam --remove clean_data/exclusions/sample.exclusions --extract clean_data/indep_snps/independent-0.05.snps --make-bed --out clean_data/hapmap_pca/dataset/${NAME}_HapMap3_pruned

After which I still get the same error.

I perform --flip-scan on my dataset, but I just get NA for NEGSNPS for all SNPs. I think it is because I have a dataset that just has controls and not cases to compare against? Regardless, can someone point me in the direction of what I can do next to investigate this problem?

Thanks.

snp software-error plink

1 answer

It sounds like you have some triallelic SNPs, which are not supported by PLINK 1's data format.

After your second merge fails, instead of using --flip [.missnp file] on your hapmap dataset, use --exclude [.missnp file] on BOTH datasets and then reattempt the merge. It is also worthwhile to note the size of the second .missnp file relative to the first: if it is much smaller, you definitely need to run --flip-scan afterwards, but if the first --flip didn't help much, there might not actually be any strand issues.

(The discussion at https://www.cog-genomics.org/plink2/data#merge3 includes a sample --flip-scan command line, which you would run after the post-exclude merge has succeeded.)

It looks like that the dataset was encoded as 1234 instead of ACGT, which is what my HapMap dataset was encoded as, so the --bmerge threw an error.

Thanks anyways!

Log in to answer this question.