To my knowledge, the two most popular software packages for analyzing population genetic data (Arlequin and Genepop) do not calculate allelic richness, which is corrected for sample size.
Fstat does, but I am interested to find a program I can run in either command line or R. Any suggestions for this?
Hi Sean,
Have you tried ARES (http://onlinelibrary.wiley.com/doi/10.1111/j.1471-8286.2007.01705.x/abstract) or HP-Rare (http://www.montana.edu/kalinowski/KalinowskiReprints/2005_HP-Rare_MolecularEcologyNotes.pdf)? ARES is an R add-on: ftp://ftp.uni-bayreuth.de/pub/math/statlib/R/CRAN/src/contrib/Descriptions/ARES.html
Thanks Christophe. Unfortunately, to my knowledge FSTAT has no command line or batch option, and I have many files to analyze.
Yep right, FSTAT has no command line or batch option, then ARES as suggested by Heini might be a solution... I'm not aware of any other at the moment....
Just for currency in this thread; this r package will do so too. It also accounts for sample size. http://cran.r-project.org/web/packages/PopGenReport/PopGenReport.pdf
As a noob R user I found PopGenReport as the most user friendly and the output is very neat and clear. I had 17507 loci and it worked perfectly on my dataset.
It would be nice to not just compare the magnitude of allelic richness, but to work out 95% confidence intervals to see if differences in magnitude are significant. I've just tried to do this in the R-package diveRsity, but am confused by the results:
4 populations, with rarefaction (smallest sample size: 20, most pops are around 40):
P1: 12.1
P2: 13.4
P3: 12.8
P4: 12.7
Without rarefaction but with CIs, 1000 bootstrap reps, alpha=0.05:
P1: mean 10.8, 95% CI: 9.6 - 11.8
P2: 12.0, 10.8 - 13.1
P3: 11.5, 10.4 - 12.5
P4: 11.3, 10.2 - 12.4
This is a strange result because I would have thought that allelic richness should be lower overall when populations are resampled to account for differences in number of individuals (i.e. rarefaction). But they are in fact larger!
Can anyone suggest what's going on here?
Thanks,
P.
Hi Peter,
Did you ever find out what was going on? I am trying to get allelic richness, private alleles and expected heterozygosity using rarefaction and with confidence intervals but few programs will do all of these measures. Even though you have generated confidence intervals, can you still interpret these differences when rarefaction hasn't been used?
Emma
Hi Emma, I never received an answer - but I suppose a simple solution to the rarefaction issue would be to simply reduce the number of individuals in each population to that of the smallest population.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics