In vitro embryo models supported by methods development in adjacent fields have revolutionized our understanding of embryogenesis.

Research with human embryos and embryo models, this year’s Method of the Year, can be fraught. In contrast, digital embryos could be studied, even perturbed, in computational what-happens-when experiments.

The creation of multiple stem-cell-derived models of mammalian embryogenesis is opening many new doors to study human development and brings a need for scientists to demonstrate responsible dialog over the associated ethical issues.

Two new methods enable the study of the mechanics of tissue development.

DeepMainmast is a protein structure modeling protocol for cryo-EM that combines the strengths of a deep-learning-based de novo protein main-chain-tracing approach with AlphaFold2-based structure predictions for improved performance.

Foambryo is an analysis pipeline for three-dimensional force-inference measurements in developing embryos.

TubULAR is an in toto tissue cartography method for mapping complex dynamic surfaces

Targettrack is a deep-learning-based pipeline for automatic tracking of neurons within freely moving C. elegans. Using targeted augmentation, the pipeline has a reduced need for manually annotated training data.

MISCOs are spatially arranged ventral midbrain–striatum–cortical organoids that enable investigations into the human dopaminergic system.

SEVtras is an algorithm that uses single-cell RNA sequencing data to assess small extracellular vesicle activity at droplet resolution.

Large language models are learning the language of genomics.

Developments in nanopore-based peptide detection and sequencing show promise of a breakthrough.

Spatially resolved multimodal omics offers a collective way to capture molecular information in complex tissues.

High-resolution connectomics of the human brain is the next frontier in neuroscience.

New twists on established methods and multimodal imaging are poised to bridge gaps between cellular and organismal imaging.