BLUP.AI fuses modern AI with the proven foundations of BLUP to extract digital phenotypes from images, integrate genomics and pedigree, and deliver sharper breeding decisions—faster.
Turn raw images into quantitative traits using ViTs, VAEs, and structured latent-space models—capturing complex morphology at scale.
Seamlessly combine A, G, and H relationship matrices with digital similarity matrices to power next-gen ssGBLUP.
Modern deep models extract features; BLUP delivers interpretable, statistically grounded genetic values for selection.
Estimate heritabilities, genetic correlations, and breeding values across classic traits and visual complex traits.
End-to-end workflows built for real breeding programs: APY support, HPC-ready training, and automated evaluation.
Secure handling of sensitive genetic and image data, with reproducible experiments and audit-friendly outputs.
We treat images and omics as complementary signals. Vision Transformers learn rich representations of morphology; generative models build structured latent spaces aligned with biological priors.
From those embeddings, we construct digital relationship matrices (D) and integrate them with classical pedigree (A) and genomic (G) information, enabling single-step evaluation for visual complex traits.
The result: more accurate breeding values, earlier selection, and better understanding of hidden biological architecture.
BLUP.AI is built from real quantitative-genetics pipelines and modern machine learning—validated on large breeding datasets and designed to plug into operational evaluations.