Infinite Physical Monkey: Do Deep Learning Methods Really Perform Better in Conformation Generation?

Conformation Generation is a fundamental problem in drug discovery and cheminformatics. Generally, it can be categorized into three different classes according to physical scales, i.e., micro molecule (organic), meso molecule (nano particle-like), and macro molecule (protein and nucleic acid). Organic molecule conformation generation, particularly in vacuum and protein pocket environments, is most relevant to drug design. Recently, with the development of geometric neural networks, the data-driven schemes have been successfully applied in this field, both for molecular conformation generation (in vacuum) and binding pose generation (in protein pocket). The former beats the traditional ETKDG method, while the latter achieves similar accuracy compared with the widely used molecular docking software. Although these methods have shown promising results for real-world drug design campaigns, some researchers have recently questioned whether deep learning (DL) methods perform better in molecular conformation generation via a “parameter-free” method. To our surprise, what they have designed is some kind analogous to the famous infinite monkey theorem, the monkeys that are even equipped with physics education. To discuss the feasibility of their proving, we constructed a real infinite stochastic monkey for molecular conformation generation, showing that even with a more stochastic sampler for geometry generation, the coverage of the benchmark QM-computed conformations are higher than those of most DL-based methods. By extending their physical monkey algorithm for binding pose prediction (with 2000 random samples), we also discover that the successful docking rate also achieves near-best performance among existing DL-based docking models. Thus, though their conclusions are right, their proof process needs more concern. In addition to evaluating the rationality of their algorithms and conclusions, we dig into the inspirations of infinite physical monkeys. We find that, for docking pose generation, DL-based models truly learn the interaction rules between residues and ligands, and discover an inductive bias hidden in the training of the pocket-given docking problem. The code of the proposed algorithm could be found at: https://github.com/HaotianZhangAI4Science/infinite-physical-monkey. TOC: Infinite Physical Monkey. This image was created with the assistance of DALL·E 2