maf is change when converting vcf to binary

[alan mejia]

Hi,

I have a vcf file from a whole exome sequencing, and I am converting this to binaries. Please rule out any normalization for now. 

I am using plink 1.9, and this is the code:

plink --vcf file.vcf --recode make--bed --out binary, then:

plink --bfile binary --maf

in vcf, maf for this rs1650697 SNP is 325/516=0.63. 

In Plink1.9, maf for the same SNP is 0.37

However, many other SNPs keep the same maf in vcf or plink.

Any thoughts?

Thanks

[Christopher Chang]

plink 1.x generally does not preserve allele order; see https://www.cog-genomics.org/plink/1.9/data#ax_allele .  plink 2.0 does not have the same problem, and should generally be used over 1.9 when working with VCF data.