Modified Algorithm for Phylogenetic Tree Construction using Ant Colony Optimization

Constructing Phylogenetic tree from DNA of species is an important problem in Bioinformatics. These trees are used to study the evolutionary relationship between species and to study which species are more closely related or distantly related. Ant colony optimization (ACO) paradigm is one of the best solutions through meta heuristics solutions which can be used to construct phylogenetic trees. Some researches present solutions using ACO, better than mostly used methods like Neighbor-Joining, UPGMA, etc. In this research we propose a modified ACO algorithm which can perform better than conventional ACO approaches in more divergent species by abstracting some features in developed ACO solutions. We construct distance matrix calculating Kolmogorov Complexity as distance between species from complete mitochondrial genome without applying Multiple Sequence Alignment (MSA). Then we obtain optimal sequence of combinations of species applying modified ACO algorithm. Using a new scenario, we construct phylogenetic tree from that sequence of combinations and test with two types of datasets. First, we tuned the model using twenty species of mammalians dataset. As second type ten divergent datasets are obtained from different divergent levels of real benchmark trees which are used to evaluate our method. Our results show that our proposed method performs better than conventional methods for divergent species. As well as it gives results in less processing time than the conventional ACO algorithm.