EmbryoHG

Predict the phenotype of embryos at the highest defensible resolution.

A Heru initiative to push embryo phenotype prediction across complex polygenic traits, built on the Haeckel dataset.

Dataset
Genotype + phenotype + extended phenotype
Model
Polygenic calibration + uncertainty
Output
Embryo phenotype distributions
DatasetModelPrediction
The dataset

A dataset Haeckel alone can build.

Haeckel connects genotype, phenotype, and extended phenotype inside one consent-governed substrate. At scale, that becomes a world-unique dataset for calibrating embryo-level prediction with ancestry, family structure, environment, and uncertainty in view.

The work

Polygenic trait calibration

Benchmark PRS methods, ensemble weights, ancestry correction, and confidence intervals trait by trait.

Embryo-level simulation

Model sibling distributions through Mendelian inheritance, recombination, parental genotypes, and phenotype priors.

Phenotype ontologies

Connect observable traits, biomarkers, morphology, cognition, disease risk, and longitudinal outcomes into one schema.

Validation

Use family data, held-out cohorts, and prospective follow-up to measure what prediction can and cannot claim.

Who should join

A project for domain experts.

Statistical geneticists
Embryologists and fertility clinicians
Human phenotype researchers
Bioinformaticians and model evaluators
Discuss collaboration

Research only. Not clinical advice. Any embryo work must be supervised and lawful where performed.