Decagon (Bio-decagon)

Bio-decagon 是一个用于多种药物副作用识别问题的数据集，该问题被框架为两个节点类型的两层多模态图/网络中的多关系链接预测问题：药物和蛋白质。蛋白质-蛋白质相互作用网络描述蛋白质之间的关系。药物-药物相互作用网络包含 964 种不同类型的边缘（每种副作用类型一个），并描述了哪些药物对导致了哪些副作用。最后，药物-蛋白质链接描述了给定药物靶向的蛋白质。链接不同数据库使用的实体词汇后的最终网络有 645 个药物和 19,085 个蛋白质节点，由 715,612 个蛋白质-蛋白质、4,651,131 个药物-药物和 18,596 个药物-蛋白质边缘连接。Bio-decagon is a dataset for the multi-drug side effect identification task, which is framed as a multi-relational link prediction problem in a two-layer multimodal graph/network with two node types: drugs and proteins. The protein-protein interaction network describes the relationships between proteins. The drug-drug interaction network includes 964 distinct edge types (one for each side effect type), and delineates which drug pairs induce which side effects. Finally, drug-protein links describe the proteins targeted by a given drug. After aligning entity vocabularies across different databases, the final network has 645 drug nodes and 19,085 protein nodes, connected by 715,612 protein-protein, 4,651,131 drug-drug, and 18,596 drug-protein edges.

搜集汇总

背景与挑战

背景概述

Bio-decagon数据集旨在识别药物副作用，通过构建一个多模态图网络来建模药物和蛋白质之间的关系。该网络包含645种药物和19,085个蛋白质节点，并通过数百万个边缘连接，包括蛋白质-蛋白质、药物-药物和药物-蛋白质相互作用。