A2 different between PLINK .bim file and .frq file?

Question

[curious]

test.vcf:

##fileformat=VCFv4.0
#CHROM  POS ID  REF ALT QUAL    FILTER  INFO    FORMAT  NA00001 NA00002 NA00003 NA00004 NA00005
19  111 .   A   C   .   .   .   GT  1/1 1/1 1/1 0/1 0/1

I run:

plink \
--vcf test.vcf \
--out test

plink \
--bfile test \
--freq \
--out test

my .bim file

19  .   0   111 C   A

my .frq file

 CHR  SNP   A1   A2          MAF  NCHROBS
  19    .    A    C          0.2       10

If the sixth column of the .bim file is supposed to be A2, why does it not agree with the .frq file?

Answer 1

[chrchang523]

See https://www.cog-genomics.org/plink/1.9/data#ax_allele . If you're trying to keep REF/ALT alleles straight, you probably want to use plink 2.0 instead of 1.9 when possible.

[curious]

Thank you, this was just something I noticed was happening and was more just curious from a broad perspective about how this happens.