Researchers at Children’s Medical Research Institute (CMRI) have developed a new framework to help scientists assess how accurately laboratory models resemble aspects of early human development.

Alternative intro (not using early human development straight up): When scientists use laboratory models for research, how do they know they can trust what they’re seeing?

Stem-cell-derived models known as blastoids are designed to resemble the human blastocyst – a structure formed during the earliest stages of development. These models offer researchers an important way to investigate early developmental biology, an area that can otherwise be difficult to study.

However, understanding how closely different blastoid models reflect what happens during early human development is critical to interpreting the research conducted with them.

CMRI researchers have now created a detailed reference map of gene activity during early human development, bringing together data from human blastocysts at different stages of development.

The team then used this reference to systematically compare different blastoid models, assessing how accurately they recapitulated key cell types and developmental processes seen during early human development.

Their analysis, published in Cell Systems, showed the models varied in how closely they reflected embryo-derived cell states, with some models performing better than others for particular cell types.

Associate Professor Pengyi Yang from CMRI’s Computational Systems Biology Unit and The University of Sydney said understanding both the strengths and limitations of these models is important as the field continues to develop.

“Stem-cell-derived blastoid models offer a promising way to investigate aspects of early human development, but it is essential to understand what they model accurately and where their limitations remain.

“We created an integrated transcriptomic reference of early human development and a multidimensional framework for comparing blastoid models with embryo-derived cell states. This provides researchers with a more objective basis for evaluating and improving these models, and for ensuring that biological conclusions are matched to what each model can reliably represent.”

The framework provides a useful benchmark for the field, helping scientists better understand the strengths and limitations of different blastoid models and improve how these models are developed and used in future research.

Read the research: https://doi.org/10.1016/j.cels.2026.101738