22 August 2024 in Pattern Recognition
19 January 2022 in IEEE/ACM Transactions on Computational Biology and Bioinformatics
Recently, deep learning has become the mainstream methodology for Compound-Protein Interaction (CPI) prediction. However, the existing compound-protein feature extraction methods have some issues that limit their performance. First, graph networks are widely used for structural compound feature extraction, but the chemical properties of a compound depend on functional groups rather than graphic structure. Besides, the existing methods lack capabilities in extracting rich and discriminative protein features. Last, the compound-protein features are usually simply combined for CPI prediction, without considering information redundancy and effective feature mining. To address the above issues, we propose a novel CPInformer method. Specifically, we extract heterogeneous compound features, including structural graph features and functional class fingerprints, to reduce prediction errors caused by similar structural compounds. Then, we combine local and global features using dense connections to obtain multi-scale protein features. Last, we apply ProbSparse self-attention to protein features, under the guidance of compound features, to eliminate information redundancy, and to improve the accuracy of CPInformer. More importantly, the proposed method identifies the activated local regions that link a CPI, providing a good visualisation for the CPI state. The results obtained on five benchmarks demonstrate the merits and superiority of CPInformer over the state-of-the-art approaches.
10 September 2024 in IEEE/ACM Transactions on Computational Biology and Bioinformatics
Recently, mask-fill-based 3D Molecular Generation (MG) methods have become very popular in virtual drug design. However, the existing MG methods ignore the chemical properties of atoms and contain inappropriate atomic position training data, which limits their generation capability. To mitigate the above issues, this paper presents a novel mask-fill-based 3D molecule generation model driven by atomic chemical properties (APMG). Specifically, we construct a new attention-MPNN-based encoder and introduce the electronic information into atom representations to enrich chemical properties. Also, a multi-functional classifier is designed to predict the electronic information of each generated atom, guiding the type prediction of elements and bonds. By design, the proposed method uses the chemical properties of atoms and their correlations for high-quality molecule generation. Second, to optimize the atomic position training data, we propose a novel atomic training position generation approach using the Chi-Square distribution. We evaluate our APMG method on the CrossDocked dataset and visualize the docking states of the pockets and generated molecules. The obtained results demonstrate the superiority and merits of APMG over the state-of-the-art approaches. The dataset and codes will be available on the project homepage: https://github.com/JU-HuaY/APMG
12 December 2024 in Research Square
Graph neural networks (GNN) have shown strong performance in link prediction tasks. However, it is susceptible to higher latency due to the trivial correlation of data in its neighborhood, which poses a challenge for its practical applica- tion. In contrast, although Multi-layer Perceptron (MLP) performs poorly, it has a shorter inference time and is more flexible in practical applications. We uti- lize a distillation model to combine the powerful inference capabilities of GNN with the inference effciency of MLP. Distillation models usually use a predefined distance function to quantify the differences between teacher-student networks, but this cannot be well applied to various complex scenarios. In addition, the limited node information severely affects the learning ability of MLP. Therefore, to cope with these problems. Firstly, we propose an Adversarial Generative Dis- criminator (AGD), which trains the discriminators and generators against each other to adaptively detect and reduce the differences. Secondly, we also propose the Feature Aggregation Module (FAM) to help the MLP obtain suffcient fea- ture information before distillation starts. In the experiments, it is shown that our approach can achieve good results in link prediction tasks, outperforming the baseline model Linkless Prediction (LLP) and maintaining a good inference speed on eight datasets in two different settings∗ . ∗The code on https://github.com/lwuen/LPVAKD.git