DeePhy: A Deep Learning Model to Reconstruct Phylogenetic Tree from Unaligned Nucleotide Sequences

---

DeePhy is a deep learning framework to infer triplet tree topology from a three sequences. The code is written in Python 3.6 and PyTorch 1.9.

Input:

This program executes by taking three nucleotide sequences (FASTA format) as a single file. To compute the accuracy, it can take the correct the tree topology as one-hot encoded format where siblings are represented as 0 and outgroup is represented as 1.

Output:

This program infer triplet topology as a class level. In Newick format, (A,(B,C)) topology is represented as Class 0, whereas, (B,(A,C)) and (C,(A,B)) topologies are representated as Class 1 and 2, respectively.

Download executable (version 1.0)

DeePhy.zip

For any queries, please contact:

• Aritra Mahapatra
  Department of Computer Science and Engineering
  Indian Institute of Technology Kharagpur
  Email: aritra DOT mhp AT iitkgp DOT ac DOT in
• Jayanta Mukherjee
  Department of Computer Science and Engineering
  Indian Institute of Technology Kharagpur
  Email: jay AT cse DOT iitkgp DOT ac DOT in