DockGen-E dataset

该数据集名为DockGen-E，包含了一系列蛋白质-配体复合物，这些复合物展现了在自然界中罕见的蛋白质中发现的不同的结合口袋，并且每个结合口袋都与一种独特的生物学功能相关联。由于这些结合口袋的新颖性，该数据集对于基于深度学习的方法来说颇具挑战性。其规模涉及具有独特生物学功能的各异结合口袋，而所面临的任务是蛋白质-配体对接以及构象改造。

This dataset, named DockGen-E, consists of a series of protein-ligand complexes featuring distinct binding pockets rarely found in naturally occurring proteins, with each binding pocket associated with a unique biological function. Owing to the novelty of these binding pockets, this dataset poses considerable challenges for deep learning-based approaches. Covering a diverse set of binding pockets with distinct biological functions, the targeted tasks for this dataset are protein-ligand docking and conformational remodeling.