Fu's FS in Arlequin and DnaSP

Question

[helena.batalha]

Hello,

I was wondering if anyone ever tried calculating Fu's FS statistic in both Arlequin and DnaSP? I did, but I get different values. I used both programs to calculate Tajima's D and Fu's FS, with the same dataset, and I get exactly the same values for Tajima's D but different values for Fu's FS. Does anyone know if the different programs calculate Fu's FS differently? I looked it up on the available documentation but couldn't find much information.

Many thanks

[ursulajaros11]

Hi Helena,

Did you finally solve the problem?

Me, I am faced with the same issue: different results for Fu's Fs in Arlequin (extremely negative ones, which are higly significant in all populations...but the result seems inplausible) and DNAsp (moderately negative and rarely significant).

Would be so interesting for me, what the reason for this discrepancy is..

Best,
Ursula

Answer 1

[katharina-kniesz]

Hey Helena & Ursula,

I had the same problem and did not find any answer. So I tried different things and figured out that DNASP is automatically cutting out the positions with missing data in your alignment (for example caused by shorter sequences). So your analysed alignment in the end has not the full length. Arlequin is analysing the full alignment. To receive the same results as calculated with DNASP you need to tell Arlequin to exclude the missing data. I am creating my arelquin files with DNASP. So before saving as arlequin file you can set the option that you want to analyse just the positions with information for all sequences :)

Best, Kati