Скачать или смотреть Mapping the Dance of Life: How Proteins REALLY Move

In this video, I break down the key ideas from our paper:
“Protein Structure–Function Relationship: A Kernel PCA Approach for Reaction Coordinate Identification.”
Understanding how protein structure leads to function is one of the central goals in computational biology and biophysics. But identifying the reaction coordinate—the low-dimensional pathway that governs functional transitions—is notoriously difficult.
This paper introduces a data-driven solution using Kernel PCA, a nonlinear dimensionality-reduction technique that uncovers hidden collective variables directly from molecular dynamics trajectories. Unlike traditional linear PCA, this method captures nonlinear conformational changes, allowing for a more accurate characterization of:
Protein activation pathways
Energy landscape structure
Functional transitions and intermediate states
Key residues contributing to the reaction coordinate
📌 In the video, I cover:
Why reaction coordinate identification is challenging in proteins
How Kernel PCA works and why it helps
Applications to protein dynamics
Examples of discovered pathways and structural features
How this method complements MD simulations and free-energy calculations
If you're interested in computational biophysics, MD simulations, machine learning for proteins, or reaction coordinate discovery, this video provides a clear and intuitive explanation.
Paper: https://pubs.acs.org/doi/full/10.1021...
www.baratilab.com