Minor allele count report with --freq count in PLINK2
1
1
Entering edit mode
14 months ago
L_to_the_m ▴ 10

Hi, I have a dataset of 451 diploid humans in PLINK2. When trying to get the count report of the minor alleles in PLINK2 with this code:

plink2 --pfile file --freq counts


I will get a table like this:

CHROM   ID  REF ALT ALT_CTS OBS_CT
1   AX-13191280__rs3131972  G   A   169 902
1   AX-32225497__rs3131962  G   A   139 900
1   AX-11214939__rs12562034 G   A   84  902
1   AX-32355233__rs12131377 C   G   59  894
1   AX-32359895__rs61768212 G   C   452 898
1   AX-32365087__rs11240779 A   G   222 900
1   AX-32366075__rs57181708 A   G   90  898
1   AX-40293871__rs4970383  C   A   222 898
1   AX-32432755__rs4970382  T   C   352 900
1   AX-40308003__rs950122   G   C   183 902
1   AX-32459995__rs13303222 G   A   136 900
1   AX-32461903__rs6657440  T   C   330 902
1   AX-32488633__rs4970459  C   T   192 902
1   AX-32519659__rs74047407 G   A   209 900
1   AX-40357457__rs1110052  T   G   249 902
1   AX-32571027__rs2272757  A   G   306 900
1   AX-32572119__rs35471880 G   A   67  894
1   AX-32598061__rs13303065 T   C   307 900
1   AX-40388697__rs13303010 A   G   86  900


But sometimes I get a Alternate allele count which is for example ALT_CTS= 452 (as in line 5). How can this be since I only have 451 diploid individuals? Is it then then still the minor allele?

0
Entering edit mode

Couldn't one sample have more than one alteration?

0
Entering edit mode

This makes sense, and then the number is the sum of all minor alleles? And in total this result in a bigger number?

0
Entering edit mode

That is what I would assume. I would definitely look into it closer. I've had a similar problem before but the issue was more clear where ALT looked more like C,A instead of just C

3
Entering edit mode
14 months ago

ALT alleles are not guaranteed to be minor.

0
Entering edit mode

Ok, is there a way to only get the minor alleles?

1
Entering edit mode

You can use --maj-ref force --make-pgen to swap all major alleles to REF, and then run --freq counts on the resulting output.