Mathematical Insights into Protein Architecture: Persistent Homology and Machine Learning Applied to the Flagellar Motor

We present a machine learning approach that leverages persistent homology to classify bacterial flagellar motors into two functional states: rotated and stalled. By embedding protein structural data into a topological framework, we extract multiscale features from filtered simplicial complexes constructed over atomic coordinates. These topological invariants, specifically persistence diagrams and barcodes, capture critical geometric and connectivity patterns that correlate with motor function. The extracted features are vectorized and integrated into a machine learning pipeline that includes dimensionality reduction and supervised classification. Applied to a curated dataset of experimentally characterized flagellar motors from diverse bacterial species, our model demonstrates high classification accuracy and robustness to structural variation. This approach highlights the power of topological data analysis in revealing functionally relevant patterns beyond the reach of traditional geometric descriptors, offering a novel computational tool for protein function prediction.

💡 Research Summary

The manuscript presents an integrated computational framework that combines persistent homology—a tool from topological data analysis (TDA)—with modern machine learning techniques to classify the functional state of bacterial flagellar motors. The authors begin by highlighting the central role of three‑dimensional protein structure in determining biological activity and point out the limitations of traditional geometric descriptors such as root‑mean‑square deviation (RMSD), sequence alignment, and energy‑based scoring functions. While these methods capture pairwise distances or global alignment, they often miss higher‑order connectivity patterns (loops, voids, and multi‑scale networks) that can be crucial for function.

To address this gap, the paper constructs a filtered simplicial complex from atomic coordinates of the flagellar motor. Starting with the set of vertices (atoms), edges are added according to a distance threshold, and finally 2‑simplices (triangles) are introduced to fill in local surface patches. This filtration yields a nested sequence of complexes K₀ ⊂ K₁ ⊂ K₂ ⊂ K₃. For each level, chain groups Cₚ(Kᵢ) and boundary operators ∂ₚᵢ are defined, leading to homology groups Hₚ(Kᵢ) = Zₚ(Kᵢ)/Bₚ(Kᵢ). Inclusion maps between filtration levels induce persistence maps ϕᵖ_{i,j}: Hₚ(Kᵢ) → Hₚ(Kⱼ). The authors formalize the entire persistence module as a graded ℤ