Fitness (SGE)

Predict continuous functional effects for assayed variants in endogenous-locus saturation genome editing screens, using the gene, assay mechanism, and local exon sequence.

published gene panels each contain 50 variants from one exon and exactly 100 unmarked flanking bases on each side. Version-2 panels sample five score-space bins across the full eligible allele population, with sparse-bin slots redistributed. The primary score is mean within-gene Spearman correlation; mean Pearson correlation reports numerical agreement, valid-output rate reports strict JSON compliance, and invalid completed outputs contribute zero while remaining identifiable as format failures.

Spearman measures ordering within each gene, not cross-assay calibration. SGE effects depend on the cellular system, engineered background, selection, timing, and treatment; they are not clinical classifications. One exon window omits distant gene and splice context, while score-space sampling does not reproduce the natural variant distribution.

Source data come from published saturation genome editing score sets in MaveDB, cited as Rubin et al. (2025).