CNNTrees

CNNTrees classifies sets of phylogenetic trees by applying convolutional neural networks to Robinson–Foulds (RF) distance matrices to identify clusters representing distinct evolutionary models.

Key Features:

• Phylogenetic Tree Classification: Uses convolutional neural networks to classify multiple phylogenetic trees and analyze heterogeneous sets containing overlapping species.
• RF Distance Matrix Analysis: Predicts the number of clusters within a Robinson–Foulds distance matrix to identify distinct evolutionary relationships among phylogenetic trees.
• Cluster-Based Evolutionary Modeling: Detects groups of trees representing different evolutionary histories influenced by processes such as horizontal gene transfer or recombination.
• Alternative Supertree Construction: Supports generation of alternative supertrees by grouping compatible phylogenetic trees and reducing topological conflicts.

Scientific Applications:

• Phylogenetic Analysis: Enables classification and comparison of multiple phylogenetic trees derived from genes or species datasets.
• Evolutionary Biology: Supports inference of evolutionary relationships and exploration of heterogeneous evolutionary histories in Tree of Life studies.
• Comparative Genomics: Assists in identifying evolutionary patterns across genes or species with differing phylogenetic signals.

Methodology:

CNNTrees computes Robinson–Foulds distance matrices from sets of phylogenetic trees and applies convolutional neural networks to these matrices to predict cluster structure and classify trees representing distinct evolutionary models.