Fig. 1: The architecture overview of DTIAM. | Nature Communications

Fig. 1: The architecture overview of DTIAM.

From: DTIAM: a unified framework for predicting drug-target interactions, binding affinities and drug mechanisms

Fig. 1

The framework mainly consists of three modules. A The drug molecular pre-training module. The module segments the molecular graph into several substructures and learns its representation through three self-supervised models from massive amounts of label-free data. B The target protein pre-training module. The module uses Transformer attention maps to learn the representations and contacts of proteins based on unsupervised language modeling from large amounts of protein sequence data. C The downstream drug-target prediction module. The module incorporates drug and target representation and predicts DTI, DTA, and MoA via an automated machine learning model.

Back to article page