VIDraft/pxr-challenge-method
收藏资源简介:
这是一个用于孕烷X受体(PXR)激动剂活性预测的数据集,专注于计算药物发现中的分子性质预测任务。数据集包含总计7,039个分子,来源于OpenADMET PXR盲测挑战赛,具体包括:官方挑战训练数据(4,139个分子)、官方反筛选数据(2,647个分子)以及公开的第1阶段模拟集(253个分子)。数据以SMILES字符串形式提供,pEC50值作为目标变量,用于预测PXR活性。该数据集用于训练和评估XGBoost集成模型,以支持ADMET(吸收、分布、代谢、排泄和毒性)属性评估。
This is a dataset for Pregnane X Receptor (PXR) agonist activity prediction, focusing on molecular property prediction in computational drug discovery. It contains a total of 7,039 molecules sourced from the OpenADMET PXR Blind Challenge, including: official challenge training data (4,139 molecules), official counter-assay data (2,647 molecules), and the publicly released Phase 1 analog set (253 molecules). The data is provided in SMILES strings with pEC50 values as target variables for PXR activity prediction. This dataset is used for training and evaluating XGBoost ensemble models to support ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) property assessment.




