Research Highlight

AI-accelerated genome-informed photosynthetic design, prediction, and optimization

Published:
Updated:
Visit the project page

Motivation: 

A biotechnology revolution is needed to increase crop productivity and energy abundance by enhancing photosynthetic efficiency. Conventional experimental and modeling approaches are too slow and fragmented to support this revolution.
 

Approach: 

Innovative AI models that link plant genomics to traits that enable photosynthetic efficiency and are trained on multimodal genomic to phenomic datasets are essential to inform photosynthetic system design, prediction, and optimization.
 

Impacts: 

  • Developed an AI foundation modeling framework linking genomics to phenomics.
  • Collected large datasets of genomics and photosynthetic phenomics and are making them AI ready.
  • Developed a prototype data lakehouse supporting genomic to phenomic linking and AI model training.
  • Evaluated and expanded DNA language models (PlantCAD2 and AlphaGenome) and have shown promising capabilities for modeling the genome of Populus trichocarpa, an important DOE energy crop.


S&T Challenge:  This GPTgp project supports the Genesis Mission Challenge “Scaling the Biotechnology Revolution”.
 

Science Area

Genesis Mission