The abstract is read at the publisher; the summary is JClub's.
A CNN trained on 25,080 maize images learned deep phenotypes that distinguish low- vs high-nitrogen responses with 96.7% accuracy and reveal NUE-linked genetic variation.
- Why it matters: Conventional plant measurements miss much of the phenotypic and genetic variation tied to nitrogen-use efficiency, limiting reliable proxies for breeding.
- What they did: The model’s learned deep phenotypes were used for genetics, yielding 523 significant loci versus 21 from conventional phenotypes, and candidate genes were interpreted across CNN layers.
- The result: Functional testing of Liguleless2 (LG2) showed lg2 mutants have enhanced root architecture and higher N-uptake efficiency, and field trials of 200 hybrids linked beneficial alleles to +18 g ear weight per plot under low-N.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.