Protein docking refinement by convex underestimation in the low-dimensional subspace of encounter complexes

Abstract

We propose a novel stochastic global optimization algorithm with applications to the refinement stage of protein docking prediction methods. Our approach can process conformations sampled from multiple clusters, each roughly corresponding to a different binding energy funnel. These clusters are obtained using a density-based clustering method. In each cluster, we identify a smooth “permissive” subspace which avoids high-energy barriers and then underestimate the binding energy function using general convex polynomials in this subspace. We use the underestimator to bias sampling towards its global minimum. Sampling and subspace underestimation are repeated several times and the conformations sampled at the last iteration form a refined ensemble. We report computational results on a comprehensive benchmark of 224 protein complexes, establishing that our refined ensemble significantly improves the quality of the conformations of the original set given to the algorithm. We also devise a method to enhance the ensemble from which near-native models are selected.

Refinement of protein‐protein complexes in contact map space with metadynamics simulations

“…Importantly, as also recently discussed by several groups in the protein docking field, 45 our results underline that the identification of improved quality snapshots, from thousands of solutions, remains one of the most challenging tasks for successful protein-protein complex refinement. The explored combination of FES and ZRANK in a simple weighted additive scoring function, CS α , although showing some encouraging results, did not yield significant success at improving this outcome.…”

A benchmarked in-silico driven pipeline capable of complete trifecta protein analysis

“…Conventional methods of NMR and X-Ray Crystallography are expensive and not readily available to the masses. Also, the structure of short half-life proteins cannot be predicted using these methods [1]. The current pipeline successfully utilizes the two main methods of protein modelling available, namely, homology and ab-initio modelling [42].…”

“…In the scientific endeavor of understanding the mechanisms of an organism's biology, the study of proteins, a key element in a number of cellular activities is vital. Proteins play both structural and functional roles within the organization of a cell [1]. The formation of a protein's characteristic and functionally related 3D structure begins from a simple but unique amino acid sequence [2].…”

“…For the determination of a protein's tertiary structure, a number of experimental procedures exist today; X-ray crystallography and Nuclear Magnetic Resonance are two of the most commonly used pathways today. But the primary shortcomings of these methods are that they lack efficiency at the same time being highly costly [1]. This disables a major portion of the scientific community from conducting comprehensive studies in relation to a protein's structure and function primarily due to the inaccessibility to such specialized equipment and the cost.…”

Controlled‐advancement rigid‐body optimization of nanosystems

“…However, the authors noticed unstable behavior of the method in several cases, particularly, when monomers in molecular complex approach too close to each other, they are pushed out, thus, deteriorating interaction interface. The latter problem is related to the pitfalls of optimization over the rugged energy landscape, which has many local minima and maxima of various shapes . Another important problem is that different energy functions form different energy landscapes given the same conformational space, thus, energy‐based methods must be specifically tuned with respect to the energy function.…”