We have mtDNA sequences (cyt b 600 bp) of fish and want to determine the 'connectivity' between populations using Arlequin, Network, DnaSP, Mega software packages. Also, need to identify unique haplotypes within populations. However, we are in need of some instructions to ensure accuracy. Many thanx!!
Gerhard, I found that the easiest way to calculate the population parameters you are interested in is to use the package Pegas in R. It is equipped with excellent step-by-step examples and MUCH more user-friendly then say Alequin where preparation of input file is pain in the brain. Even if you did not ever use R before! Just checked it with local students ;). http://r-project.org. Just one warning: drawing of minimal spanning trees is not as nice in R as in Network.
Dear Gerhard,
in the text of your question you listed a number of valuable computer programs that may calculate the parameters you are interested in. As regards their interpretation, I cannot only suggest 1) to study the theory behind each parameter (many computer manuals are quite clear for that) and 2) to read as much as possible scientific articles on experimental investigations similar to your work. I believe this is the only way to avoid misinterpretations of experimental results. Also, there is a third way for lazy guys, namely to analyse data together with an expert... However, too easy and quick shortcuts should be avoided in serious science.
Dear Gerhard,
To overcome misunderstandings, you should comprehend the principle and the implemented mathematical formula in assessing the Fst and AMOVA parameters using each genetic software. You have already cited the adequate programs that enable this (I personally prefer Arlequin package!!). Yes, I guess the next step might be going pertinently through equations, then doing some good readings about this (I would suggest, e.g: Ecological Genetics: Design, Analysis and Application by Andrew Lowe, Stephen Harris, Paul Ashton, Arlequin manual, you can also try FSTAT, Genalex and Genetix manuals, they explain very well the Fst even if they use nuclear markers such as SSR rather than mtDNA).
Good luck!
You can find some basic notes on population genetics (heterozygosity, Fst, genetic distances etc ...) in this on-line tutorial: http://www.uwyo.edu/dbmcd/molmark/.
I hope this helps!
Gerhard, estimation of population differentiation is always hampered by sampling. So a critical aspect to make a decision about using F statistics is whether the sampling is representative enough. From that standpoint, if your idea is identifying some connectivity between populations, a quick way to do it could be doing genome wide analysis using pool DNA samples for each group. That way you can quickly detect markers that are shared among your groups. Then you can track the specific frequencies of those shared regions in each population and use Fst and other pop parameters to find out whether there is a pattern of introgression or migration that created the similarity at those loci.
Gerhard,
Shane Lavery has an excellent introduction to Fst (linked below) that I use for teaching, and Chris Bird has an excellent review of the subject (also linked) with Peter Smouse among others, who was an author on the original AMOVA paper. As Amal says you could use these to get better acquainted with the subject.
My own research and that of others has left me with the opinion that Fst (of any flavor- Phist, Rst, Gst) is not so good for connectivity analysis, especially in marine species, so use and interpret with care. Robin Waples had an influential article about this (below). It is often depressed by high heterozygosity (nucleotide diversity in the case of mtDNA) to where it is not significantly different from zero and thus without information to test hypotheses. However, for much the same reason (high genetic variability), a coalescent analysis such as Migrate can get more information for you than it could for many terrestrial species. See my linked article below for an example.
http://www.library.auckland.ac.nz/subject-guides/bio/pdfs/733Pop-g-stats2.pdf
Article Separating the wheat from the chaff: Patterns of genetic dif...
Article Coalescent and biophysical models of stepping-stone gene flo...
Chapter Detecting and measuring genetic differentiation
Yes, I agree with Eric that fst is not so reliable in connectivity analysis, especially at low sample size. AMOVA is more elaborate and probably more trustworthy. And don't forget to replicate! Remember that comparing populations within the same ecosystem is probably pseudo-replication.
Andrew :-)
- Badges
- Science topic
- Similar topics
- Filing