bed/bim/fam format and phase

[Matthew Maher]

I'm attempting to use a tool that requests:
"Phased imputation data in PLINK format"   (bed/bim/fam)

I think of plink2 format (pgen/pvar/psam) as explicitly tracking phased-versus-not, while plink1 format (bed/bim/fam) does not.   But would I be correct to think that, while it doesn't explicit track it,  the (bed/bim/fam) format can maintain phase as long as I'm careful about how the data are loaded and/or manipulated?  i.e. is the format internally storing two ordered calls for each sample/variant? 

So I believe (okay, I hope) that if I have a phased VCF file and do e.g.:
plink2 --vcf <VCFFILE> --make-bed --out <FILESET>
then the BED file is still phased, even though it's not explicitly tracked.

Is my understanding correct?
And I suspect I should avoid plink 1.9 in this context?

Thanks for any info, and thanks for PLINK*

[Chris Chang]

That’s incorrect.  Each .bed file entry is just a count in {0, 1, 2, missing}.

Now, for a single sample, you can try to arrange things such that the A1 allele always represents one haplotype and the A2 allele represents the other.  But this doesn’t scale to more than one sample, and you have to be careful to never scramble allele order.

--
You received this message because you are subscribed to the Google Groups "plink2-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plink2-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/plink2-users/aa8794db-ea81-497a-a50a-ad2ad3012e87n%40googlegroups.com.

[Matthew Maher]

thank you for that rapid clarification.