Druglike molecule datasets for drug discovery

<strong>Background</strong><br> Trnasformer-based AI models have shown outstanding performance in identifying druggable candidate molecules. In most cases, models are trained on a massive amount of database of molecular information to capture the latent meaning of a given molecule. However, the desirable properties of candidate molecules include the feasibility of synthesizing them, low toxicity, and high druggability. In this study, we injected prior knowledge of the desirable properties of molecules during the training process. <strong>Methods</strong><br> Using the PubChem database (100 M), we filtered druglike molecules based on the quantity of drug-likeliness (QED) score and the Pfizer rule. With this dataset of drug-like molecules, we trained both the molecular representation model (chemBERTa) and the molecular generation models (MolGPT). The molecular representation model was evaluated by fine-tuning the results on the MoleculeNet benchmark datasets, and the molecular generation model was evaluated based on the generated samples (10 K). <strong>Results</strong><br> Training with druglike molecules enabled the generation of molecules with desirable properties without any conditioning. Although the molecular representation learning model was not remarkable, however, its performance in predicting clinical toxicology exceeded that of conventional molecular representation models. <strong>Conclusion</strong><br> By training based on a dataset of druglike molecules, our approach enables molecular representation models to predict clinical toxicity more precisely. Furthermore, it enables the molecule generation model to generate molecules with desirable druglike properties without any conditional generation procedures.<br> ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- import pickle with open("druglike_molecules_QED.pkl", "rb") as f: data = pickle.load(f)**背景** 基于Transformer (Transformer)的人工智能模型在可药用候选分子的识别任务中展现出优异性能。多数场景下，此类模型会在大规模分子信息数据库上进行训练，以捕获给定分子的潜在语义表征。然而，候选分子的理想属性应涵盖合成可行性、低毒性与高成药性。本研究在训练过程中注入了分子理想属性的先验知识。 **方法** 本研究使用PubChem数据库（100 M，即1亿条数据），基于定量药物相似性评分 (QED, Quantitative Estimate of Drug-likeness) 与辉瑞规则筛选类药分子。基于该类药分子数据集，我们分别训练了分子表征模型chemBERTa与分子生成模型MolGPT。分子表征模型通过在MoleculeNet基准数据集上微调完成性能评估，分子生成模型则基于生成的1万条样本（10 K）进行性能验证。 **结果** 使用类药分子进行训练，可在无需任何条件约束的情况下生成具备理想属性的分子。尽管分子表征学习模型的基础表现并不突出，但其在临床毒理学预测任务中的性能超越了传统分子表征模型。 **结论** 基于类药分子数据集开展训练的研究方法，可使分子表征模型更精准地预测临床毒性。此外，该方法还可使分子生成模型在无需执行任何条件生成流程的前提下，生成具备理想类药属性的分子。 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- import pickle with open("druglike_molecules_QED.pkl", "rb") as f: data = pickle.load(f)