This Perspective discusses the application of algorithmic methods throughout the preclinical phases of drug discovery to accelerate initial hit discovery, mechanism-of-action elucidation and chemical property optimization.

RAS proteins, central drivers of cancer, appeared ‘undruggable’ for almost 30 years. Here we provide a personal perspective on the effort leading to our initial report of KRAS^G12C inhibitors in 2013, and the decade of discoveries that followed.

Boron is a common element found in various minerals; however, it is not used in life-creating machinery in nature. In a new study, boron enzymes have been created by introducing a boronic acid-containing non-canonical amino acid into an artificial enzyme scaffold. Further development via directed evolution enabled the selection of new-to-nature stereoselective chemistry.

The inaugural CRISPR-based drug Casgevy has been approved by several medical agencies, with other CRISPR-based therapies currently in clinical trials. Although there are technological hurdles to overcome, chemical biology has a vital role in developing recent breakthroughs in base editing, prime editing and epigenetic editing into future treatments.

The incorporation of nitrogen in steroidal glycoalkaloids is hypothesized to occur through a transamination reaction. Here, the authors show that GLYCOALKALOID METABOLISM12 appears to evolve from the canonical γ-aminobutyric acid transaminases and directs the biosynthesis of nitrogen-containing steroidal metabolites in Solanum plants.

Caliendo and Vitu et al. developed a platform for endogenous gene expression monitoring and conditional actuator activation, by integrating noncoding RNAs downstream of the polyadenylation signal without altering endogenous gene coding sequences.

Superoxide oxidizes the iron–sulfur cluster in the DNA demethylase ROS1 and extensively participates in the establishment of active DNA demethylation in plants to maintain stem cell fate.

By enriching productive mutational paths, a Kemp eliminase that speeds up proton abstraction >10⁸-fold was developed in only five evolution rounds. Recombining it with a variant differing by 29 substitutions revealed the underlying fitness landscape.

How a lasso cyclase ties a lasso peptide into its characteristic knot has remained poorly understood. Here the authors identify key molecular interactions that guide lasso peptide folding and cyclase substrate tolerance to inform cyclase engineering for expanded lasso peptide diversity.

Cryo-electron microscopy and biochemical analysis reveal the activation mechanism of protein arginine kinase McsB by its activator McsA for protein quality control under stress in Gram-positive bacteria.

Computational methods for calculating a protein structure from a given amino acid sequence have revolutionized both our understanding of structural biology and the prediction of protein-binding compounds. This issue features several pieces that explore machine learning approaches for protein structure prediction, benchmarking and evaluation of model quality, and how machine learning algorithms can be used in the drug discovery process.