Question

Absolute Allele Frequency Difference

0

Entering edit mode

11.5 years ago

meganathan.pr ▴ 10

Hi, I would like to calculate absolute allele frequency differences (delta) for each SNPs of two populations (using 1000 genomes vcf data). I could not find the correct formula for this. Could anyone tell me how to do?

population genetics • 7.4k views

ADD COMMENT • link updated 10.4 years ago by Biostar 20 • written 11.5 years ago by meganathan.pr ▴ 10

score 3 · Answer 1 · 2013-05-21

3

Entering edit mode

11.5 years ago

Giovanni M Dall'Olio 28k

The delta Allele Frequency is just the difference of the frequencies of the allele in the two populations. However, to calculate it, you must first choose which allele to use to calculate the frequency: the derived allele. If you know which is the ancestral allele, you can calculate the delta DAF. Otherwise, the deltaMAF is fine for most cases.

absoluted delta Derived Allele Frequency (deltaDAF) -> frequency of the most recent allele (not the Ancestral Allele)
absolute delta Minor Allele Frequency (deltaMAF) -> frequency of the allele that has the minor frequency in a reference population (e.g. in CEU).

1000 Genomes already provides an annotation on which is the minor allele (see the VCF file).

ADD COMMENT • link 11.5 years ago by Giovanni M Dall'Olio 28k

0

Entering edit mode

Thanks for your answer. My main objective is to select AIMs panel for my population of interest. I have seen some articles where they have used Fst Value to preliminarily screen SNPs. However, Fst varies from 0 to 1. To know the baseline of Fst to select AIMs some authors used delta value. In this case, should I use deltaDAF or deltaMAF? Do you have any other suggestions to select the AIMs based of Fst?

ADD REPLY • link 11.5 years ago by meganathan.pr ▴ 10

0

Entering edit mode

Sorry, I don't have experience in selecting Ancestral Informative Markers. I think that the answer may depend on how reliable are the annotations on the ancestral alleles: for 1000 Genomes, you can pretty much trust the ancestral state annotations, so you can use the deltaDAF. For other datasets, where you are not sure about the ancestral state, you may want to use the deltaMAF. In any case, why don't you open a new discussion for this question? (e.g. something with the title "what is the best approach to select Ancestral Informative Markers") You would get more visibility and more answers.