Too many distinct nonstandard chromosome/contig names

[Rohan Flint]

Hi,

I am trying to convert my vcf file containing only SNP variants to plink format using the command;

plink --vcf my_file.vcf --allow-extra-chr --out my_file

The reference genome I am using contains 77390 contigs, and when I try to run this command, it returns the error 'Error: Too many distinct nonstandard chromosome/contig names.' 

I have tried using the '0' modifier for the --allow-extra-chr option but it still returns the same error.

Some examples for the contig names are 'ScfyRBE_1.HRSCAF.2' for contig 1 and 'ScfyRBE_2.HRSCAF.32' for contig 2. 

I have tried this on both PLINK v2.00a3 64-bit (17 Feb 2020) and PLINK v1.90b6.16 64-bit (17 Feb 2020) but both versions return this error.

I was wondering if there were any workarounds for this, as I am trying to perform other plink analyses on this data.

Thanks.

[Christopher Chang]

I'd recommend the following:

1. If your variants don't have distinct IDs, you probably want to use a short shell script to give them unique position-based IDs (which include contig names).

2. Once you've done that, it should be ok to use shell one-liner to zero out the chromosome column of the VCF.

[Christopher Chang]

For anyone who runs into this problem in the future: if you clone the source code from GitHub and uncomment "#define HIGH_CONTIG_BUILD" at the top of plink2_common.h, and then recompile, you'll get a plink2 build that should support up to about a million contigs; and it's straightforward for me to raise this limit for anyone that needs it.

On Monday, July 26, 2021 at 5:37:35 AM UTC-7 flyn...@gmail.com wrote: