Nati Daniel

The paper introduces a diploid genomic foundation model to represent diploid genotypes in genomic language models, improving variant discrimination in compound-heterozygous cases.

Wrap a genomic foundation model's inference methods once to score it on supervised and zero-shot GUE, VariantBenchmarks, BEND, TraitGym and ClinVar tasks; code is promised upon acceptance.

The paper uses machine learning to analyze embryo developmental timing and shows that maternally aged embryos lose temporal regulation, impacting viability.