search

JonJala / mama

MIT License
13 stars 4 forks source link

ValueError: keep_snps indices out of bounds #27

Closed ldcato closed 3 years ago

ldcato commented 3 years ago

Hi, excited about this software - congrats.

Running on RedHat linux, within the virtualenv. I am running the latest version (i.e. followed your page advice and ran git pull to update). Tutorial data runs smoothly. I have this error using other data (see log attached below).

Summary of data: 4 ancestry groups. iid_ances_file and snp_ances_file formatted as per the wiki. There are no duplicated IDs in the snp_ances_file or in the genotype data. There are "multi-allelic variants" i.e. variants at the same position but represented as sepeate variants (for example in: VCF =

CHROM POS REF ALT

2 10304 A T,C PLINK Bfile = 2 10304 A T 2 10304 A C ) But they all have different RSIDs so I doubt this is causing the problem?

One last addition: I am running mama_ldscores.py per chromosome but my snp_ances_file is just one entity chr1 -> chr22. Again, I am doubting this is causing the issue?

Would greatly appreciate your help. I'm a bit stuck as to what the problem is. Thanks! Liam

chr1_EUR_AFR_SAR_THAI.log

ggoldman1 commented 3 years ago

Hey Liam,

Almost certainly the issue is that your genome-wide snp_ances file does not correspond to chromosome-level bim file. Having rsID's in your snp_ances file that aren't in your bim file will cause an indexing error. Can you make 22 of the snp_ances files, one for each chromosome, and let us know if that fixes it?

Grant

ldcato commented 3 years ago

Hi Grant! Ah thank you, sorry about that! Currently running - it's father than its ever gotten before! Thanks for your help!!

ggoldman1 commented 3 years ago

Hey Liam, has this issue been resolved?

Grant