WelQrate
收藏资源简介:
WelQrate数据集是由范德堡大学的研究团队精心策划的高质量小分子药物发现基准数据集,涵盖了9个数据集,跨越5种治疗目标类别。数据集的大小从约66K到300K不等,包含高度不平衡的活动标签,以反映低命中率的现实情况。创建过程中,通过多层次的筛选和领域专家的严格预处理,如PAINS过滤,确保数据的高质量。WelQrate数据集的应用领域主要集中在小分子药物发现,旨在通过提供高质量的数据集和标准化的评估框架,推动AI模型在实际药物发现中的应用,解决现有数据集质量不高的问题。
The WelQrate dataset is a high-quality small-molecule drug discovery benchmark dataset carefully curated by a research team from Vanderbilt University. It encompasses 9 datasets spanning 5 therapeutic target categories, with the size of each dataset ranging from approximately 66K to 300K. The dataset contains highly imbalanced activity labels to reflect the realistic scenario of low hit rates in drug discovery. During its creation, multi-level screening and strict preprocessing conducted by domain experts (such as PAINS filtering) were implemented to ensure its high data quality. The WelQrate dataset is primarily applied in the field of small-molecule drug discovery, aiming to promote the practical application of AI models in real-world drug discovery by providing high-quality datasets and a standardized evaluation framework, thereby addressing the problem of insufficient quality of existing datasets.

- 1WelQrate: Defining the Gold Standard in Small Molecule Drug Discovery Benchmarking范德堡大学 · 2024年



