For this lab, we worked with our alignment of COI sequences that included our DNA fish barcode sequences and sequences downloaded from NCBI via Geneious.
The first part of the lab was cleaning up our alignment, ensuring the lengths of all the sequences were lined up with one another. In looking at the first 20 columns of the alignment, 9 polymorphic sites were detected.
Using this alignment, the iModelTest2 program was used to determine the best model of molecular evolution. Based on the Akaike Information Criterion (AIC) method, the best model was TVM+G, with a score of 11471.30. The best model using Bayesian Information Criterion (BIC) was HKY+G, which scored 11807.94. BIC and AIC did not choose the same molecular model.
Next, MrBayes (Bayesian Analysis) was performed using our alignment. The following parameters were used: substitution model was HKY; Outgroup was KJ825841; Chain Length was set at 10,100; Subsample Frequency was set at 200; Burn-In Length was set at 100; Heated Chains was set at 4.
However, these results are indication that the analysis was not run for an adequate amount of time. Ideally, the first graph should depict a bell-shaped curve and the second graph should depict a “fuzzy caterpillar.”
Another run using MrBayes was performed with the following adjustments made: Chain Length was set to 1,100,000 and Burn-in Length was set to 100,000. The results of the posterior distribution, trace, and tree with support values are shown below.
Following Bayesian Analysis was Maximum Likelihood using RAxML. A consensus tree was built with the following parameter: Support Threshold was set at 50%.Finally, a maximum likelihood analysis with bootstrapping was performed using PHYML to generate our final tree. The molecular evolution model used for this method was HKY85 and the Branch Support and Number of Bootstraps were set at “bootstrap” and 100, respectively. In addition, the outgroup (KJ825841) was established as the root.