Application of deep learning techniques for de novo generation of molecules, termed as inverse molecular design, has been gaining enormous traction in drug design. The representation of molecules in SMILES notation as a string of characters enables the usage of state of the art models in natural language processing, such as Transformers, for molecular design in general. Inspired by generative pre-training (GPT) models that have been shown to be successful in generating meaningful text, we train a transformer-decoder on the next token prediction task using masked self-attention for the generation of druglike molecules in this study. We show that our model, MolGPT, performs on par with other previously proposed modern machine learning frameworks for molecular generation in terms of generating valid, unique, and novel molecules. Furthermore, we demonstrate that the model can be trained conditionally to control multiple properties of the generated molecules. We also show that the model can be used to generate molecules with desired scaffolds as well as desired molecular properties by conditioning the generation on scaffold SMILES strings of desired scaffolds and property values. Using saliency maps, we highlight the interpretability of the generative process of the model.

Viraj Bagal

Rishal Aggarwal

P. K. Vinod

U. Deva Priyakumar

Journal of Chemical Information and Modeling

JOURNAL OF CHEMICAL INFORMATION AND MODELING 发表论文报告化学信息学和分子建模领域的新方法和/或重要应用。具体主题包括化学数据库的表示和基于计算机的搜索、分子建模、新材料、催化剂或配体的计算机辅助分子设计、化学软件的新计算方法或有效算法的开发，以及包括生物分析在内的生物制药化学活动和其他与药物发现相关的问题。精明的化学家、计算机科学家和信息专家关注本月的富有洞察力的研究、编程创新和软件评论，以跟上这个完整的、多学科领域的进步。作为订阅者，您将了解数据库搜索系统、图论在化学问题中的应用、子结构搜索系统、模式识别和聚类、化学和物理数据分析、分子建模、图形和自然语言界面、文献计量和引文分析以及综合设计和反应数据库。

The Journal of Chemical Information and Modeling publishes papers that report new methods and significant applications in the fields of cheminformatics and molecular modeling. Specific topics include the representation of chemical databases and computer-based searches, molecular modeling, computer-aided molecular design of new materials, catalysts, or ligands, the development of new computational methods or efficient algorithms for chemical software, as well as research related to biopharmaceutical chemistry activities and other drug discovery issues. This journal attracts chemists, computer scientists, and information specialists who are interested in insightful research, programming innovations, and software reviews to keep up with advancements in this comprehensive, multidisciplinary field. As a subscriber, you will learn about database search systems, the application of graph theory to chemical problems, substructure search systems, pattern recognition and clustering, chemical and physical data analysis, molecular modeling, graphical and natural language interfaces, bibliometrics and citation analysis, as well as integrated design and reaction databases.

《Journal of Chemical Information and Modeling》发表关于化学信息学和分子建模领域的新方法和重要应用的论文。具体主题包括化学数据库的表示与计算机搜索、分子建模、新材料、催化剂或配体的计算机辅助分子设计、化学软件的新计算方法或有效算法的开发，以及生物制药化学活动和药物发现相关问题的研究。该期刊吸引了化学家、计算机科学家和信息专家关注本月的研究成果、编程创新和软件评测，以跟上这一多学科领域的进展。作为订阅者，您将了解数据库搜索系统、图论在化学问题中的应用、子结构搜索系统、模式识别与聚类、化学和物理数据分析、分子建模、图形和自然语言界面、文献计量与引文分析，以及综合设计与反应数据库等内容。

Category	Quartile
化学	2
化学, 药物化学	2
化学, 化学综合	2
化学, 计算机信息系统	3
化学, 计算机跨学科应用	3

MolGPT: Molecular Generation Using a Transformer-Decoder Model

ivySCI AI Smartly Parses PDF, Answers Researchers' Questions, and Helps You Understand Papers in Seconds

Journal Info

Category	Quartile
COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS	1