We have just completed our first meeting and the next meeting will be Wednesday October 17th at 11am. We will be receiving more information here or by e-mail from Nadir who will present Parsimony theory and using the Software PAUP. We look forward to having everyone there.
In the discussion today, we discussed the paper "Computer programs for population genetics data analysis: a survival guide" by Excoffier and Heckel (2006, Nature Reviews- Genetics) and talked about the directions for the group. We will meet every two weeks on Wednesdays at 11am and will make decisions about what will be discussed as we go along. As well, we reserve the right to use our meeting times or the Wednesdays in between as opportunities to discuss our work (past or present) for our own benefit!
Today:
- Genepop (population genetics) and fasta (sequence data) should be our go-to formats, using Convert as our vehicle for generating the formats for software requiring other formats
- Violaine talked about an MS-tools utility that converts data in excel to genepop format.
- Population sizes!
- Anahi has a Parsimony book available in her office for the time being
- I have added a comment forum on the sidebar. Please use it! If it is not adequate, I will look for something else. It does not have e-mail notification.
Program information:
Structure- the new version of Structure (2.2) now can accommodate dominant markers
DNAsp- the use of moving windows a powerful tool
SPAGeDi- (not pronounced like "spaghetti"?) can compute distances with methods other than Fst by isolating one population relative to others
IM- capable of discriminating between migration and isolation
GDA- finding private alleles in populations
Genepop- released version 4 last summer (after the article was published) so is a good starting point for your population data
Arlequin- use for linkage disequilibrium estimates as there are no assumptions of diploidy
MULTIPURPOSE PROGRAMS in table 2- with the exception of MSA, all need large populations of greater than 20 individuals to function properly- if you have fewer, individual based programs might be the answer.
RESISTANT to small or heterogeneous population sizes: d(mu)2, Nei distances and Cavalli-Sforza distances.
Information:
dominant and codominant markers and the programs that are capable of dealing with them
Admixture vs. no admixture- definitions and assumptions
Possible future topics:
- bayesian methods
- assumptions of models and theory behind them
- population sizes and the use of different programs and methods
- how to work with haploid data
- Geneland, BAPS and Structure- relative strengths and weaknesses and when to use each one (i.e. all have assumption of HWE, while BayesAss+ does not)
- should we always minimize assumptions?
Wednesday, October 3, 2007
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment