The Year of Open Science has highlighted the importance of sharing the code associated with peer-reviewed manuscripts. We at Nature Computational Science provide support — via policies and implementations within our submission system — to facilitate this task.

Autoencoders are versatile tools for molecular informatics with the opportunity for advancing molecule and drug design. In this Review, the authors highlight the active areas of development in the field and explore the challenges that need to be addressed moving forward.

The capability of predicting stable materials is important to further accelerate the discovery of novel materials. In this Review, the authors discuss recent developments in machine learning techniques for assessing the stability of materials and highlight the opportunities in further advancing the field.

LOVAMAP is an analysis software that accurately identifies 3D pores in packed particle systems by exploiting information about the particle configuration as a basis for segmentation. Using the software, the authors were able to uncover striking relationships between particle and pore properties.

VSSR-MC is a Markov chain method based on virtual adsorption sites that interfaces with a neural network force field to provide fast, accurate and comprehensive sampling of material surfaces.

Zhi Liu et al. develop a method to measure disparities in reporting delays in urban crowdsourcing systems, uncovering socioeconomic disparities and providing actionable insights for interventions that enhance the efficiency and equity of city services.

A graph-based contrastive learning framework, LACL, is proposed for geometric domain-agnostic prediction of molecular properties to alleviate the need for molecular geometry relaxation, enabling large-scale inference scenarios.

LOVAMAP is an analysis software that accurately identifies 3D pores in packed particle systems by exploiting information about the particle configuration as a basis for segmentation. Using the software, the authors were able to uncover striking relationships between particle and pore properties.

Designing accessible, interoperable and reusable software for applying deep learning to the study of gene regulation has been a challenge in genomics research. EUGENe is a toolkit that addresses this gap and streamlines end-to-end analyses.

A hybrid machine learning–physics model is developed that reduces simulation cost by two orders of magnitude while retaining high ab initio accuracy, to predict free-energy transition states for hydrogen combustion reactions.

The Year of Open Science has highlighted the importance of sharing the code associated with peer-reviewed manuscripts. We at Nature Computational Science provide support — via policies and implementations within our submission system — to facilitate this task.

Artificial intelligence (AI) drives innovation across society, economies and science. We argue for the importance of building AI technology according to open-source principles to foster accessibility, collaboration, responsibility and interoperability.