Vcftools option --non-ref-ac

[MurielGB]

Hello !

Can someone explain the --non-ref-ac option ?

I don't manage to get it

What happened if I do:

vcftools --vcf input.vcf --non-ref-ac 1 --out outout --recode

Thanks !!!

Muriel

[mastal]

the following quote is from the manual page:

http://vcftools.sourceforge.net/man_latest.html#SITE%20FILTERING%20OPTIONS

--non-ref-ac <integer>


'Include only sites with all Non-Reference (ALT) Allele Frequencies (af) or Counts (ac) within the range specified, and including the specified value. The default options require all alleles to meet the specified criteria, whereas the options appended with "any" require only one allele to meet the criteria. The Allele frequency is defined as the number of times an allele appears over all individuals at that site, divided by the total number of non-missing alleles at that site.'

[MurielGB]

Thanks Mastral, I checked the manual before asking the quetsion
But I don't understand what this means...
Sorry.

[mastal]

I was about to edit my reply.

the mannual has a list of options:

--non-ref-af <float>
--max-non-ref-af <float>
--non-ref-ac <integer>
--max-non-ref-ac <integer>

--non-ref-af-any <float>
--max-non-ref-af-any <float>
--non-ref-ac-any <integer>
--max-non-ref-ac-any <integer>

Reading some of the previous items in the manual,
I think '--non-ref-ac' is the minimum alternate allele count,

and if you want a range of allowable allele counts from min to max for the alternate allele,
then you can also set the '--max-non-ref-ac' option.

[MurielGB]

So, if I understand well, this command "--non-ref-ac 1" allows to remove sites where there is no alternative allele at all ?

That's what I want to do, since I filtered out some individuals in my vcf, I want to get rid of invariant sites that remain (sites where there was alternative allele but after removing individuals, some may not be variant anymore).

[SNPsaurus]

You could also use the --maf, which is from 0-1. It is the minimum allele frequency in the population. If you set --maf > 0 but less than 1/(num individuals), it should exclude all alternate alleles that are no longer in the population.