rdk/p2rank

P2Rank is a high-utility, infrastructure-grade tool within the structural biology and drug discovery niche. With nearly a decade of history and over 400 stars (a significant number for specialized bioinformatics software), it has established itself as a standard utility for pocket detection. Its defensibility stems from 'scientific lock-in': it is extensively cited in academic literature and integrated into established virtual screening pipelines. While newer deep learning approaches (e.g., DeepSite, Kalasanty, or Graph Neural Networks) offer potentially higher accuracy, P2Rank's speed and lack of heavy GPU requirements keep it relevant for high-throughput screening. Frontier labs like DeepMind (AlphaFold) or EvolutionaryScale (ESM) focus on the broader problem of structure prediction and protein design; while AlphaFold 3 can predict ligands, specialized pocket detection tools remain necessary for downstream docking and medicinal chemistry workflows. The primary risk is displacement by newer 'geometric deep learning' models that can better capture long-range interactions, but P2Rank's role as a robust, easy-to-use CLI tool provides a meaningful moat against purely algorithmic challengers.