Collective Variable for Metadynamics Derived From AlphaFold Output

Abstract

AlphaFold is a neural network–based tool for the prediction of 3D structures of proteins. In CASP14, a blind structure prediction challenge, it performed significantly better than other competitors, making it the best available structure prediction tool. One of the outputs of AlphaFold is the probability profile of residue–residue distances. This makes it possible to score any conformation of the studied protein to express its compliance with the AlphaFold model. Here, we show how this score can be used to drive protein folding simulation by metadynamics and parallel tempering metadynamics. Using parallel tempering metadynamics, we simulated the folding of a mini-protein Trp-cage and β hairpin and predicted their folding equilibria. We observe the potential of the AlphaFold-based collective variable in applications beyond structure prediction, such as in structure refinement or prediction of the outcome of a mutation.

Selecting Collective Variables and Free-Energy Methods for Peptide Translocation across Membranes

“…The reconstruction of the obtained 2D free-energy surface was performed with R plugin metadynminer . Metadynminer was also used to identify the minimal path between the adsorbed (A) and inserted (I) states and to integrate out the second dimension for the construction of a 1D free-energy profile from a surface.…”

Computational investigation on lipid bilayer disruption induced by amphiphilic Janus nanoparticles: combined effect of Janus balance and charged lipid concentration

“…Furthermore, the value of CV should reflect the state of the simulated system, including metastable states. 43 The FES can be constructed in the space spanned by those CVs. The bias potential V bias ( S , t ) at time t can be written as: 42 where ω is the Gaussian height, which is controlled by the deposition stride, S i is one of the predefined CVs, and σ i is the Gaussian width for i th CV.…”

“…Furthermore, the value of CV should reflect the state of the simulated system, including metastable states. 43 The FES can be constructed in the space spanned by those CVs. The bias potential V bias (S,t ) at time t can be written as: 42 V bias ðS; tÞ ¼…”

Collective Variable-Based Enhanced Sampling: From Human Learning to Machine Learning

“…Specialized machine learning frameworks, such as AlphaFold2, have achieved remarkable success in predicting biomolecular structures. Leveraging the information captured by these purpose-built artificial intelligence systems to extract CVs appears to be a promising approach . In future work, XAI may be employed to elucidate the movement modes discerned by these specialized machine learning frameworks .…”