Multi-GPU-based detection of protein cavities using critical points

- CriticalFinder is the first multi-GPU-based cavity detection algorithm.
- CriticalFinder is the first surface-based cavity detection algorithm that produces a meaningful, coarse cavity segmentation.
- CriticalFinder sustains on the theory of critical points (i.e., Morse theory).

Protein cavities are specific regions on the protein surface where ligands (small molecules) may bind. Such cavities are putative binding sites of proteins for ligands. Usually, cavities correspond to voids, pockets, and depressions of molecular surfaces. The location of such cavities is important to better understand protein functions, as needed in, for example, structure-based drug design. This article introduces a geometric method to detecting cavities on the molecular surface based on the theory of critical points. The method, called CriticalFinder, differs from other surface-based methods found in the literature because it directly uses the curvature of the scalar field (or function) that represents the molecular surface, instead of evaluating the curvature of the Connolly function over the molecular surface. To evaluate the accuracy of CriticalFinder, we compare it to other seven geometric methods (i.e., LIGSITECS, GHECOM, ConCavity, POCASA, SURFNET, PASS, and Fpocket). The benchmark results show that CriticalFinder outperforms those methods in terms of accuracy. In addition, the performance analysis of the GPU implementation of CriticalFinder in terms of time consumption and memory space occupancy was carried out.