BackOnTrack: Evolutionary-scale protein language models uncover beneficial variants in a sorghum bicolor diversity panel

This dataset accompanies a study investigating the use of protein language models for identifying functionally relevant genetic variation in Sorghum bicolor. It includes evolutionary conservation scores derived from the ESM2 protein language model, SIFT scores, degeneracy annotations, and inferred ancestral allele states for nonsynonymous variants. These resources were used in analyses linking protein-level evolutionary constraint to estimates of the distribution of fitness effects and phenotypic variation in sorghum.