Explainable AI identifies recombination and chromatin environment as key predictors of subgenome evolution in maize and Brassica.

In maize and Brassica rapa allopolyploids, eXplainable AI identifies recombination rate as the most influential, highly interconnected predictor of subgenome identity.

• Why it matters: Biased gene loss and regulatory rewiring after polyploidy are well known, but the specific genomic and epigenomic drivers of biased fractionation among subgenomes are unclear.
• What they did: The study built supervised machine learning classifiers using 60 genomic/epigenomic features in maize and 45 in Brassica rapa, then used XAI to rank feature contributions and map covariation networks.
• The result: XAI consistently ranked chromosome location, transposon density, and chromatin context—accessible chromatin proximity and active histone marks H3K9ac and H3K27ac—among top contributors, with recombination rate as the central network node.

AI summary of the abstract · may contain errors · not medical advice · read the paper.