Correct command line syntax for plink2

Question

[dec986]

Hello,

I'm running PLINK v2.00a2.3LM 64-bit Intel (24 Jan 2020) and I cannot get the command line syntax correct for LD pruning.

I've been reading https://avikarn.com/2019-07-30-prunning/ and https://www.cog-genomics.org/plink/2.0/ld but am confused about what I've read there. There are no examples in the manual pages, which isn't very helpful.

First, I convert vcf files thus vcftools --gzvcf $vcf --plink --out $dir/$id

I've tried

~/Scripts/plink2 --noweb --file plink/MDMNFYMQ --indep 50 5 2 --out MDMNFYMQ.indep which I got from the tutorial

but this gives the error: Error: Unrecognized flag ('--file'). which is strange, so I tried

~/Scripts/plink2 --noweb plink/MDMNFYMQ --indep-pairwise 50 5 2 --out MDMNFYMQ.indep

but then I get

Error: Invalid --ndep-pairwise r^2 threshold '2'. For more info, try "plink2 --help <flag name>" or "plink2 --help | more

I'm totally stuck.

How can I run LD pruning with plink2?

Answer 1

[chrchang523]

1. Use plink 1.9 for this. --file and --indep are not implemented in plink 2.0 yet.
2. With both plink 1.9 and plink 2.0, VCF files should be imported directly with --vcf. vcftools --plink is very inefficient and lossy in comparison.

[dec986]

I'm having a really hard time getting all of the plink input files ready.

/Scripts/plink1.9/plink --vcf genetic-data/ZZFNMDMF.vcf.gz --out ZZFNMDMF --make-founders --make-bed

generates .nosex, .log, .bed, .fam, .bin files. However, I cannot generate the ped file that is also required:

703404669@bioitutil2:~/covid_study2065$ ~/Scripts/plink1.9/plink --vcf genetic-data/ZZFNMDMF.vcf.gz --out ZZFNMDMF --recode compound-genotypes
PLINK v1.90b6.18 64-bit (16 Jun 2020)          www.cog-genomics.org/plink/1.9/
(C) 2005-2020 Shaun Purcell, Christopher Chang   GNU General Public License v3
Logging to ZZFNMDMF.log.
Options in effect:
  --out ZZFNMDMF
  --recode compound-genotypes
  --vcf genetic-data/ZZFNMDMF.vcf.gz

128738 MB RAM detected; reserving 64369 MB for main workspace.
--vcf: ZZFNMDMF-temporary.bed + ZZFNMDMF-temporary.bim + ZZFNMDMF-temporary.fam
written.
661125 variants loaded from .bim file.
1 person (0 males, 0 females, 1 ambiguous) loaded from .fam.
Ambiguous sex ID written to ZZFNMDMF.nosex .
Using 1 thread (no multithreaded calculations invoked).
Before main variant filters, 1 founder and 0 nonfounders present.
Calculating allele frequencies... done.
Total genotyping rate is 0.925025.
661125 variants and 1 person pass filters and QC.
Note: No phenotypes present.
Error: --recode compound-genotypes cannot be used with multi-character allele
names.

Is there an alternative to Plink that doesn't require so much error-prone preparation?

[chrchang523]

1. No .ped file is required (and .ped files should almost always be avoided in 2020 since they're horribly inefficient; this is why --file has been deliberately excluded from all plink 2.0 alpha builds). Use --bfile instead of --file to load the .bed + .bim + .fam fileset.
2. With that said, LD pruning (and several other standard QC operations) requires a dataset with at least ~50 samples. It is not applicable to a single-person dataset.