University of Sydney researchers have developed a powerful new way to test how closely lab-grown biological models resemble real human embryos, finding that while some models perform well, none yet fully capture the complexity of early human development.
Research on embryo models derived from stem cells, known as blastoids, offers scientists a way to study the biological events that underpin fertility, pregnancy success and early human development without relying on donated human embryos.
Models derived from stem cells could help answer questions that have long been difficult to investigate because of technical and ethical constraints. These stem-cell-derived structures are research models, not actual human embryos. Current models cannot develop into a human embryo.
However, just because something looks like an embryo, does that mean it is behaving like one biologically?
Researchers at the University of Sydney have developed a computational framework that helps answer that question. Their study systematically evaluated four leading human blastoid-generation methods and found substantial differences in how faithfully they reproduce the cell types and developmental processes seen in natural human embryos.
Published in Cell Systems , the research creates one of the most comprehensive reference maps of early human embryo development and uses it to benchmark the biological accuracy of stem-cell-derived embryo models.
Lead author Associate Professor Pengyi Yang , ARC Future Fellow in the School of Mathematics and Statistics and Unit Head of Computational Systems Biology at the Children’s Medical Research Institute, said the work provides researchers with a more objective way to understand the strengths and limitations of embryo models.
"Human embryo models have enormous potential for studying the earliest days of an embryo’s development, but there has been no consistent way to assess how accurately these reflect real human development," said Associate Professor Yang, who also leads the Trans-Regulatory Biology group at the Charles Perkins Centre.
"Our framework allows researchers to compare these models against a detailed biological reference and determine which cell types and developmental processes are faithfully reproduced, and which are not.”
HOW THEY MADE THE MAP
The team combined and harmonised more than 14,000 single-cell transcriptomes (the set of RNA molecules in a cell) from human embryos spanning key stages of their development. This allowed them to create a reference map of how cells normally differentiate and organise themselves during the days immediately before and after implantation.
They then compared that reference against four widely used blastoid-generation protocols developed by international research groups. Rather than assessing whether the models simply resembled embryos under a microscope, the researchers examined their molecular identities, developmental timing, lineage structure and other biological characteristics.
The results showed that some blastoid models reproduced all three major cell lineages of a natural human blastocyst, or the early embryo, relatively well, while others failed to accurately represent certain cell types or contained large numbers of cells that could not be confidently matched to any known embryonic state. No single model perfectly replicated a natural human blastocyst.
Associate Professor Yang said the findings highlight both the promise and current limitations of embryo models.
“The encouraging finding is that some models capture important aspects of early embryonic development relatively well, although each model has limitations,” he said.
“But our study also shows that current models are not biologically equivalent to real human embryos, and researchers need to be careful about the conclusions they draw from them.”
By providing a standardised benchmark, the researchers hope future models can be improved more rapidly and evaluated more rigorously.
“If we're going to use these systems to answer important biological questions, we first need to know what they can reliably tell us,” Associate Professor Yang said.
“Our work provides a roadmap for improving embryo models and ensuring scientific claims remain grounded in what the models can actually support.”
The researchers have made their reference datasets and benchmarking tools publicly available, enabling scientists worldwide to test new embryo models against the same standards.
DOWNLOAD photos of the researchers and a copy of the research at this link .
INTERVIEWS
Associate Professor Pengyi Yang | pengyi.yang@sydney.edu.au | +61 452 536 773
Please contact Associate Professor Yang by email or SMS first to arrange interview time. Or contact the media adviser below.
MEDIA ENQUIRIES
Marcus Strom | marcus.strom@sydney.edu.au | +61 474 269 459
Outside of work hours: please call +61 2 8627 0246 (directs to a mobile number) or email media.office@sydney.edu.au .
RESEARCH
Long, S. et al ‘Systematic transcriptomic evaluation of blastoid models of early human development’ ( Cell Systems 2026) DOI: 10.1016/j.cels.2026.101738
DECLARATION
The authors declare no competing interests.
The work was supported by a National Health and Medical Research Council Investigator Grant and a Metcalf Prize for Stem Cell Research to Pengyi Yang. Raja Jothi was supported by the Intramural Research Program of the National Institutes of Health (USA) and National Institute of Environmental Health Sciences (USA).
Cell Systems
Computational simulation/modeling
Cells
Systematic transcriptomic evaluation of blastoid models of early human development
2-Oct-2026