I finished the previous lab by conducting a BLAST search of the COI gene from 24 ray-finned fishes (Actinopterygii) and one shark (Chondrichthyes). I created alignment of these 25 COI genes, and my two samples EZ02 and EZ03.
I started this lab by cleaning up the alignment, and deleting any extra base pairs from each end. Next, I used the program jModelTest2 to find the best model for these sequences. I exported my alignment from Geneious in a relaxed Phyllip format (.phy), and loaded it into jModelTest2.
From this alignment, I computed the likelihood scores, keeping all defaults for this calculation. I then used two different methods to find the best model. I first tried the Akaike Information Criterion (AIC) method. The best resulting method from this analysis was TVM+G. Then I ran the Bayesian Information Criterion (BIC) method. The best model based on BIC was TPM2uf+G.
Next, I ran a Bayesian inference to create a tree using MrBayes in Geneious. I used ‘HKY85’ for the substitution model, and gamma for the rate variation. I selected the shark gene for the outgroup. I first used a chain length of 10,000, but later repeated the methods with a chain length of 3,000,000. I used an unconstrained branch length and shape parameter of 10.
Next, I used maximum likelihood to infer a phylogenetic tree of my aligned data set. I did this with the RAxML plugin for Geneious. I used the rapid bootstrap with rapid hill climbing method. I used the resulting file to create a consensus tree with a support threshold of 50%.
Then, I tried a different program for maximum likelihood, called PHYML.
Figure 1. Final Phylogeny
Figure 2. Phylogengy Trace
Figure 3. Parameter Estimates