AIIRA - AI Institute For Resilient Agriculture转发了
Most plant properties (phenotypes) are determined by a mix of genetics (G) and environment (E). But wouldn't it be simpler if some phenotypes were determined by genetics and others by environment? That's what Anirudha Powadi, Talukder Zaki Jubery & Michael Tross set out to do in this new preprint. They designed a new computational encoder strategy to decompose hyperspectral reflectance data from a maize diversity panel of ~700 genotypes grown in two environments into a set of genetic and environmental factors that can be used to predict whole plant traits. Anirudha Pawadi, Talukder Zaki Jubery, Michael Tross, James Schnable, Baskar Ganapathysubramanian (2024) Disentangling Genotype and Environment Specific Latent Features for Improved Trait Prediction using a Compositional Autoencoder.?https://lnkd.in/gw-kzWHD Really fun project supported by AIIRA - AI Institute For Resilient Agriculture.