Once we had the our consensus sequences from our 2 samples that worked through the PCR, we were able to determine the best model of molecular evolution. In order to do this, we used jModelTest2 to figure the best model. I downloaded jModel2 from the site given and opened a folder with a program that uses JAVA. We then went back to Geneious and exported our alignment in Phylip format (relaxed). Afterwards, we went back to JAVA and opened the exported alignment and computed the likelihood scores. Once the scores were done, we chose the best model based on some optimality criterion.Two methods were used: the Alkaike Information Criterion (AIC) and Bayesian Information Criterion (BIC). The best model shown for AIC was 11369.78 and the bet model based on BIC was 11622.87, which are the same model.
We then looked at MrBayes analysis, and we made a tree with substitution model GTR and rate variation +G (gamma). The MCMC was set to 10,100. The subsample frequency was set to 200, and the heated chains were left at 4. The priors for this Bayesian analysis were unconstrained branch length: Exponential (10), and Shape parameter (10). The distribution shown demonstrated that our analysis was too short.
We used maximum likelihood to infer a phylogenetic tree of our aligned data set with RAxML We choose our evolutionary model as rapid bootstrap with rapid hill climbing. Once the RAxML was done, we built a consensus tree with a support threshold of 50%. The clades with high support match those from the Bayesian analysis.
We also tried a different program that uses maximum likelihood, PHYML, which we ran with bootstrapping and HKY85 as out molecular model of evolution. We used the tree window to manipulate our resulting best ML tree with bootstrap support by making the outgroup the root, showing th bootstrap proportions, and making the tree easy to read.
We used the same alignment and the HKY85 model of molecular evolution to infer the best tree using MrBayes. We ran it with 3 million generations and a subsample frequency of 500. The outgroup was chose to be a ray.When the run is done, we exported the final tree with support values, the posterior distribution, and the trace.