Introduction to the themed collection on ‘AI in Medicinal Chemistry’
Jian Zhang, Ola Engkvist and Gerhard Hessler introduce the RSC Medicinal Chemistry themed collection on ‘AI in Medicinal Chemistry’.
AiZynth impact on medicinal chemistry practice at AstraZeneca
The AI retrosynthesis tool AiZynth has made positive impacts on AstraZeneca drug discovery projects. This opinion provides some examples and discusses how AI retrosynthesis fits into pharmaceutical research.
Research Article
Identification of lysosomotropism using explainable machine learning and morphological profiling cell painting data
Explainable ML was used to identify important chemical structural properties that contribute to lysosomotropism.
Extension of multi-site analogue series with potent compounds using a bidirectional transformer-based chemical language model
Shown is the extension of an analogue series with a new potent compound using a chemical language model. Substitution sites and non-hydrogen R-groups are colored in red (the log-likelihood score for the new analogue is reported in parentheses).
Geometric deep learning-guided Suzuki reaction conditions assessment for applications in medicinal chemistry
Machine learning-predicted screening plate.
Research Article
Leveraging bounded datapoints to classify molecular potency improvements
We present a novel data pre-processing approach, “DeltaClassifier”, that enables classification models to access traditionally inaccessible bounded datapoints to guide molecular optimizations by directly contrasting pairs of molecules.
Research Article
Identification of SARS-CoV-2 Mpro inhibitors through deep reinforcement learning for de novo drug design and computational chemistry approaches
A pragmatic approach to the discovery of new SARS-COV-2 Mpro inhibitors by combining generative chemistry and computational chemistry approaches.
Research Article
Can large language models predict antimicrobial peptide activity and toxicity?
The large language models GPT-3 and GTP-3.5 were challenged to predict the activity and hemolysis of antimicrobial peptides from their sequence and compared to recurrent neural networks and support vector machines.
Systematic generation and analysis of counterfactuals for compound activity predictions using multi-task models
For a kinase inhibitor correctly predicted with a multi-task machine learning model (shown on an orange background), counterfactuals with small chemical changes (shown in red) were generated that were predicted to be active against other kinases.
Research Article
Discovery of novel SOS1 inhibitors using machine learning
Machine learning enabled ligand-based virtual screening is a valuable tool in discovering effective SOS1 inhibitors.
Research Article
Deconvoluting low yield from weak potency in direct-to-biology workflows with machine learning
Augmenting direct-to-biology workflows with a new machine learning framework.
About this collection
This themed collection, guest edited by Professor Jian Zhang (Shanghai Jiao Tong), Professor Ola Engkvist (Astrazeneca) and Dr Gerhard Hessler (Sanofi), highlights the latest advances in the field of artificial intelligence applied to medicinal chemistry. The field of AI has progressed very fast during the last decade. The progress has impacted many fields including drug discovery and development. This collection covers various areas related to the application of AI for designing novel compounds of medical relevance as well as using AI for designing the synthetic routes for medicinal chemistry relevant compounds. Areas of how the design of compounds can be combined through AI with automation, information rich assays like image & transcriptomics assays and large language models like GTP4 will also be included.
New articles will be added to the collection upon publication. Please return to this page frequently to see the collection grow.