PubMed · 2026-06-30
AI systems trained on millions of plant sequences, protein structures, and field images are now decoding gene regulation, engineering proteins, scoring crop traits across thousands of plants at once, and automating literature searches that once consumed months of researcher time. Plant biology's longstanding bottleneck, too much data to interpret, is cracking open.
Five distinct application areas are already active in plant science: genomic regulatory decoding, protein engineering, visual phenotyping at breeding-population scale, cross-species cell-type annotation, and AI-powered research workflow automation
Vision foundation models can assess plant traits across entire breeding populations simultaneously, a task previously requiring manual scoring of individual plants
Single-cell foundation models can annotate cell types across multiple species, eliminating the need to build separate models for each crop