errors in running smartpca

Question

[ShirleyDai]

Hi, I've got an error when running smartpca. Anyone has ideas what was going wrong? Thanks.

$ perl smartpca.perl -i raw-GWA-data.hapmap3r2.pruned.bed -a raw-GWA-data.hapmap3r2.pruned.pedsnp -b raw-GWA-data.hapmap3r2.pruned.pedind -o raw-GWA-data.hapmap3r2.pruned.pca -p raw-GWA-data.hapmap3r2.pruned.plot -e raw-GWA-data.hapmap3r2.pruned.eval -l raw-GWA-data.hapmap3r2.pruned.log -k 2 -t 2 -w pca-populations.txt
smartpca -p raw-GWA-data.hapmap3r2.pruned.pca.par >raw-GWA-data.hapmap3r2.pruned.log
fatalx:
snps out of order and packed format.  Run convertf with pordercheck: NO
Aborted (core dumped)
ploteig -i raw-GWA-data.hapmap3r2.pruned.pca.evec -c 1:2  -p   -x  -y  -o raw-GWA-data.hapmap3r2.pruned.plot.xtxt
evec2pca.perl 2 raw-GWA-data.hapmap3r2.pruned.pca.evec raw-GWA-data.hapmap3r2.pruned.pedind raw-GWA-data.hapmap3r2.pruned.pca

Answer 1

[LauferVA]

This problem can be caused by a variety of issues, but in my experience all of them relate to the .map or .bim file.

1) you have chr 0 or chr 24-26 data in your data set. Solution: Fix with plink --exclude or plink --chr 1-23

2) you have multiple data points with a genetic distance listed as "0" found after SNPs with a higher genetic distance. Solution: remove using a command like awk '($3==0){print $2}' mysnpdata.map > snpstoexclude.txt then remove with plink.

3) you have data between chr 1-23 and genetic distances are assigned appropriately, but the snps are out of order. Solution: Fix with plink --recode.

Answer 2

[spsoisson]

I was getting the same error. I had SNPs labeled as being on chromosome 0. I went back to plink and removed those using --chr 1-23. I ran smartpca again and it worked.

[ShirleyDai]

Thanks. I will try it.