bcftools view output: afs

[CowGirl]

Hi everyone,
I've got a question about the output of bcftools view.

I'd like to call variants in 31 bam files with samtools.

Here's my command line:

samtools mpileup -A -C50 -E -D -S -uf Chr24.fa *_chr24.bam | bcftools view -cvgb - > chr24.raw.bcf 2 > chr24.log

As output, I get the following:

[bcfview] 100000 sites processed.
[afs] 0:96554.433 1:449.486 2:156.922 3:276.314 4:178.592 5:112.159 6:120.255 7:88.550 8:65.583 9:70.643 10:84.772 11:81.573 12:98.371 13:68.041 14:37.212 15:28.736 16:30.144 17:28.370 18:26.915 19:26.635 20:24.097 21:34.090 22:39.297 23:35.057 24:35.446 25:35.173 26:44.701 27:41.310 28:44.901 29:40.760 30:44.085 31:182.506 32:36.646 33:54.260 34:35.597 35:12.299 36:14.098 37:10.696 38:8.856 39:10.062 40:9.075 41:6.398 42:7.110 43:5.393 44:5.720 45:5.883 46:5.821 47:8.709 48:8.830 49:10.590 50:13.063 51:9.918 52:14.322 53:8.406 54:10.751 55:9.196 56:13.714 57:27.756 58:141.493 59:164.972 60:47.778 61:25.470 62:71.993

... and then I get different values for 200000, 300000 etc. sites processed.

What exactly do these numbers mean? I'd appreciate anyones help!

Cheers and thanks in advance

[Lluc]

As bcftools calls variants it reports the allele frequency spectrum (afs). You may be interested only in the last of these lines, when all sites have been processed. The lines starting with [afs] are a series of pairs n:m, where n goes from 0 to twice the number of (diploid) individuals in the sample (2N), and m is the expected number of variable sites where the alternative allele is present in exactly n copies out of 2N. These numbers are estimated from the genotype likelihoods and from a prior expectation of the distribution of alleles. The latter is controlled by the bcftools option -P. At the bottom of the following link you can see how to use the output of a previous run as a prior for the following in order to refine the estimate of the AFS.

http://samtools.sourceforge.net/samtools.shtml