TransformerCPI: improving compound-protein interaction prediction by sequence-based deep learning with self-attention mechanism and label reversal experiments

By Lifan Chen, Xiaoqin Tan, Dingyan Wang, Feisheng Zhong, Xiaohong Liu, Tianbiao Yang, Xiaomin Luo, Kaixian Chen, Hualiang Jiang, M. Zheng and 1 more. 2020

527 citations recorded by the source corpus.

MOTIVATION Identifying compound-protein interaction (CPI) is a crucial task in drug discovery and chemogenomics studies, and proteins without three-dimensional (3D) structure account for a large part of potential biological targets, which requires developing methods using only protein sequence information to predict CPI. However, sequence-based CPI models may face some specific pitfalls, including using inappropriate datasets, hidden ligand bias, and splitting datasets inappropriately, resulting in overestimation of their prediction performance. RESULTS To address these issues, we here constructed new datasets specific for CPI prediction, proposed a novel transformer neural network named TransformerCPI, and introduced a more rigorous label reversal experiment to test whether a model learns true interaction features. TransformerCPI achieved much improved performance on the new experiments, and it can be deconvolved to highlight important interacting regions of protein sequences and compound atoms, which may contribute chemical biology studies with useful guidance for further ligand structural optimization. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online. AVAILABILITY AND IMPLEMENTATION https://github.com/lifanchen-simm/transformerCPI.

Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

1. Jianzhu Ma
2. Jian Peng
3. Jiahua Rao
4. Huiying Zhao
5. Jiaqi Guan
6. Sheng Chen
7. Shuangjia Zheng
8. Yuedong Yang

In Materials Chemistry

1. Jianzhu Ma
2. Jian Peng
3. Jiahua Rao
4. Huiying Zhao
5. Wesley Wei Qian
6. Jiaqi Guan
7. Sheng Chen
8. Qianmu Yuan

In Molecular Biology

1. Jianzhu Ma
2. Jinbo Xu
3. Yuedong Yang
4. Jian Peng
5. Jiahua Rao
6. Huiying Zhao
7. Wesley Wei Qian
8. Jiaqi Guan

Citations in the corpus

• TANKBind: Trigonometry-Aware Neural NetworKs for Drug-Protein Binding Structure Prediction (2022, methodology): “We compare TankBind against two state-of-the-art sequencebased methods, TransformerCPI [53] and MONN [28], two complex-based methods, IGN [38] and PIGNet [54] both requiring prior knowledge of the inter-molecular structure to predict affinity, and two structure-based methods, HOLOPTOT [23] and STAMP”
• Interpretable bilinear attention network with domain adaptation improves drug–target prediction (2022, unknown): “However, Chen et al. (2020) 16 pointed out that the Human dataset had some hidden ligand bias, resulting in the correct predictions being made only based on the drug features rather than interaction patterns.”
• 3D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Prediction (2023, methodology): “…et al., 2021b) with the unsupervised features h L provided by our model, and compare it with two state-of-the-art sequence-based models TransCPI (Chen et al., 2020) and MONN (Li et al., 2020), one complex model IGN (Jiang et al., 2021), two structure-based model HOLOPROT (Somnath et al., 2021)…”
• AlphaFold2-aware protein-DNA binding site prediction using graph transformer (2022, background): “On the other hand, transformer [36] is well acknowledged as the most powerful neural network in modelling sequential data, such as natural language[37], drug SMILES [38] and protein sequence [39].”