Scaling a context-aware genomic model for condition-specific edit design in plants

CROP-1 predicts context-specific gene activity from cis-regulatory DNA across six plant species with a median per-gene Pearson correlation of 0.831, versus 0.486 for a matched sequence-only model.

• Why it matters: Selective crop edits must preserve gene roles in other tissues, stages, or conditions, but sequence-only predictors lack explicit biological-state context, limiting reliable condition-specific design.
• What they did: CROP-1 combines DNA sequence, transcription factor binding motifs, and measured expression to predict gene activity as a function of both sequence and regulatory state, then performs gradient-based sequence optimization to design cis-regulatory edits.
• The result: In soybean and tomato, edits with 1–20 nucleotide substitutions yield strong predicted target responses while producing much smaller average changes across non-target contexts, enabling computational control of where, when, and under which conditions genes are active.