Retrosynthesis prediction is a fundamental problem in organic chemistry and drug synthesis. We proposed an end-to-end deep learning model called CTsynther (Contrastive Transformer for single-step retrosynthesis prediction model) that could provide single-step retrosynthesis prediction without external reaction templates or specialized knowledge. The model introduced the concept of contrastive learning in Transformer architecture and employed a contrastive learning language representation model at the SMILES sentence level to enhance model inference by learning similarities and differences between various samples. Mixed global and local attention mechanisms allow the model to capture features and dependencies between different atoms to improve generalization. We further investigated the embedding representations of SMILES learned automatically from the model. Visualization results show that the model could effectively acquire information about identical molecules and improve prediction performance. Experiments showed that the accuracy of retrosynthesis reached 53.5% and 64.4% for with and without reaction types, respectively. The validity of the predicted reactants is improved, showing competitiveness compared with semi-template methods.

MOTIVATION: In the field of pharmacochemistry, it is a time-consuming and expensive process for the new drug development. The existing drug design methods face a significant challenge in terms of generation efficiency and quality. RESULTS: In this paper, we proposed a novel molecular generation strategy and optimization based on A2C reinforcement learning. In molecular generation strategy, we adopted transformer-DNN to retain the scaffolds advantages, while accounting for the generated molecules' similarity and internal diversity by dynamic parameter adjustment, further improving the overall quality of molecule generation. In molecular optimization, we introduced heterogeneous parallel supercomputing for large-scale molecular docking based on message passing interface communication technology to rapidly obtain bioactive information, thereby enhancing the efficiency of drug design. Experiments show that our model can generate high-quality molecules with multi-objective properties at a high generation efficiency, with effectiveness and novelty close to 100%. Moreover, we used our method to assist shandong university school of pharmacy to find several candidate drugs molecules of anti-PEDV. AVAILABILITY AND IMPLEMENTATION: The datasets involved in this method and the source code are freely available to academic users at https://github.com/wq-sunshine/MomdTDSRL.git.

With the rapid growth of virtual drug data- bases, the need for efficient molecular docking tools for large-scale screening is also growing. We have developed Vina@QNLM 2.0, a novel molecular docking system that leverages the logical processing units and computational processing arrays of heterogeneous multicore architecture processors. Compared to Vina@QNLM, the new version optimizes the docking speed without sacrificing accuracy. This greatly improves the scoring capability for large molecules (molecular weight > 500). Simultaneously, the new system provides enhanced support for applications such as reverse target finding through an improved parallel strategy. Vina@QNLM 2.0 achieves a speedup 20 times higher than that, using logical processing units only during a single docking process. Additionally, we successfully scaled the reverse target finding a task to 122,401 kernel groups with a robust scalability of 80.01%. In practice, we completed a reverse target-seeking for nine glycan molecules with 10,094 proteins within 1 hour.