VCF tools --freq output file

Question

[Inquisitive8995]

Hello,

I am using vcftools --freq to obtain allele frequencies of my VCF file consisting of 4 individuals. I used the below command to run

vcftools --vcf A.vcf --freq --out A.frq

My output file looks like below:

CHROM   POS     N_ALLELES       N_CHR   {ALLELE:FREQ}
contig_1        279875  2       0       T:-nan  C:-nan
contig_3        277244  3       0       G:-nan  A:-nan  T:-nan
contig_3        277247  2       0       C:-nan  T:-nan
contig_4        8794    2       0       A:-nan  G:-nan
contig_4        78125   2       8       G:0     A:1
contig_4        219961  2       8       G:0     C:1
contig_4        250382  2       8       T:0     C:1
contig_11       123877  2       6       T:0.166667      C:0.833333

I was unable to find the description of the ouput headers for the .frq file. Which allele comes first, Major allele or Minor alelle? What does -nan signify?

Please let me know if can find this information anywhere.

Any help would be appreciated. Thank you so much!

Answer 1

[Kevin Blighe]

Hey,

The header for the output in included in the output itself. For example:

CHROM           POS     N_ALLELES   N_CHR   {ALLELE:FREQ}
contig_11       123877  2           6       T:0.166667  C:0.833333

From this, I can see that there are 2 unique alleles (N_ALLELES) at position contig_11:123877, and these are observed across 6 total alleles (N_CHR) - these have the following frequencies:

• T:0.166667
• C:0.833333

My crude mathematics tell me that there is 1 T base, and 5 C bases.

Kevin

[Inquisitive8995]

Hi, Thanks for your reply. If I go by the calculation, can I confidently say that the first column is MAF for creating a MAF plot. What would I infer from a SNP having three alleles as seen in contig_3:277244. Do have any suggestions about it ? Thank u

[Kevin Blighe]

For the multi-allelic site, you may want to remove those, or at least split them - see my answer here: A: Remove duplicate SNPs only based on SNP ID in bcftools

Irrespective of multi-alleles or not, I am not sure, given the history of these programs, that you can have 100% confidence that the first column always relates to the minor allele. I would implement a check via awk in order to detect the minor allele and then take that.

[Inquisitive8995]

Thank you. I suppose I can use,

bcftools view --max-alleles 2

for retaining positions with only 2 alleles. Then, I will try awk and obtain just the minor allele frequency.