AncestralClust

AncestralClust clusters divergent nucleotide sequences using a phylogenetic approach that infers ancestral sequences to improve clustering accuracy and balance.

Key Features:

• Phylogenetic Tree Construction: Builds a neighbor-joining phylogenetic tree from a user-specified subset of initial sequences.
• Initial Clustering via Longest Branches: Forms initial clusters based on the longest branch lengths in the phylogenetic tree to reflect evolutionary divergence.
• Ancestral Sequence Inference: Performs multiple sequence alignment, applies midpoint rooting, and infers an ancestral sequence at each cluster root node.
• Greedy Clustering Algorithm: Assigns sequences to clusters by genetic proximity to the inferred ancestral sequence and reserves sequences whose distance exceeds the average inter-cluster distance for future iterations.
• Iterative Refinement: Repeats the clustering process until all sequences are assigned to optimize cluster accuracy and balance.

Scientific Applications:

• Homologous Gene Analysis: Clusters homologous nucleotide sequences across species for comparative and phylogenetic analyses.
• Evolutionary Biology of Divergent Taxa: Enables examination of genetic relationships among distantly related organisms.
• Sequence Evolution and Functional Divergence: Produces more accurate and evenly sized clusters to support studies of sequence evolution and functional divergence.

Methodology:

Constructs a neighbor-joining tree from an initial subset, selects initial clusters via longest branches, performs multiple sequence alignment and midpoint rooting to infer ancestral sequences at cluster roots, and applies a greedy, distance-based assignment reserving sequences exceeding the average inter-cluster distance, iterating until all sequences are assigned.