I have sequenced several molecular markers (ITS-2, 18S and 28s rRNA genes and two mitochondrial genes, cox1 and cox2) from different mosquito species and populations. I would like to analyze the relationship of these species' population with the molecular markers mentioned.
Hi Jose,
From a phylogenetic perspective, you could produce phylogenetic trees for your taxa using MrBayes (Bayesian method; http://mrbayes.sourceforge.net/) and RAxML (Maximum Likelihood method; http://www.exelixis-lab.org/). If you're dealing with closely related species, I strongly advise you to estimate phylogenetic trees for each individual locus in addition to the concatenated matrix. This may be useful to detect incongruence among data sets, which may have biological meaning.
For a more population genetics approach, you could use Network (http://www.fluxus-engineering.com/sharenet.htm) or TCS (http://darwin.uvigo.es/software/tcs.html) to produce a network of haplotypes. If you suspect recombination among the studied species or population, SplitsTree (http://www.splitstree.org/) is also an useful tool to visualize recombinant/hibdridization events. To detect population structure, you can use Structure (http://pritch.bsd.uchicago.edu/software.html).
To address more specific problems, you'll need to elaborate more.
Hope it helps,
Diogo
For a first step of analysis you may try "Splits tree" software which is a good alternative to NJ and networck methods. After that I guess that you can use coalescent based methods and the BEAST software.
Hi Jose,
From a phylogenetic perspective, you could produce phylogenetic trees for your taxa using MrBayes (Bayesian method; http://mrbayes.sourceforge.net/) and RAxML (Maximum Likelihood method; http://www.exelixis-lab.org/). If you're dealing with closely related species, I strongly advise you to estimate phylogenetic trees for each individual locus in addition to the concatenated matrix. This may be useful to detect incongruence among data sets, which may have biological meaning.
For a more population genetics approach, you could use Network (http://www.fluxus-engineering.com/sharenet.htm) or TCS (http://darwin.uvigo.es/software/tcs.html) to produce a network of haplotypes. If you suspect recombination among the studied species or population, SplitsTree (http://www.splitstree.org/) is also an useful tool to visualize recombinant/hibdridization events. To detect population structure, you can use Structure (http://pritch.bsd.uchicago.edu/software.html).
To address more specific problems, you'll need to elaborate more.
Hope it helps,
Diogo
Basically, there are four methods used in phylogenetic tree reconstruction viz. Neighbor-Joining (NJ), Maximum Parsimony (MP), Maximum Likelihood (ML) and Bayesian Inference (BI). All these four methods can be used for your data. Now, all methods will not produce same tree topology, so it is always better to use all four methods (as I have done in most of my papers). NJ is based on distance matrix, MP on shared derived characters, ML and BI on likelihood values (tree with highest likelihood value), although BI is different from ML. NJ, MP and ML are available in Phylip and BI in MrBayes. Now the advantage of MrBayes is that gaps can be used as characters after coding them as binary data, utilizing a software like Seqstate. Also, since you are using genes from nuclear as well as mitochondria, MrBayes can partition your data and by applying jModeltest, you can choose the best model (of evolution) that fits your data. MP of Phylip codes gaps as fifth character (NJ, ML and BI codes gaps as missing data). At the population level, NJ might be a better option. Hope it helps!
Irrespective of phylogenetic gene tree, you can used sequence data to compare population structure within species using AMOVAs as in Arlequin software.
To compare trees obtained from different markers you can use the Incongruent Lenght Difference (ILD) test. To compare trees obtained from distinct markers or reconstruction method (NJ, MP, ...) you can use the Shimodaira-Hasegawa test. Both test are available in PAUP.
if your dataset includes 30-50 samples of cox1 and|or cox2 sequences per population, you can access genetic population polymorphism and divergence by SNP in SMOGD, Popgene and DNaSP. See http://www.ngcrawford.com/django/jost/
Just one other thought, STRUCTURE was mentioned, to estimate population structure among accessions. It uses a maximum likelihood approach, and there is another package, ADMIXTURE, which uses a Bayesian approach to the same problem, to give you the Bayesian advantages.
- Badges
- Science topic
- Similar topics
- Biological Science
- Evolutionary Biology
- Phylogenetics