Region / Region set — conservation, age, selection and functional strata for a built-in set, an uploaded BED or a single-study workflow.
BED uses 0-based, half-open chromosome/start/end intervals. Overlapping and adjacent intervals are merged; different alleles at the same position remain distinct.
The method and denominator for the current selection appear above. Bootstrap spread, subsampling intervals and standard errors are labeled separately.
Spearman ρ between Evo2 Δ (selected model) and allele age (GEVA AgeMean_Jnt) / recent selection (SDS_Final, re-signed to this database's allele orientation), for All / LCR / non-LCR subsets. LCR = RepeatMasker repFamily present. Estimates use site subsampling, not bootstrap resampling: each of 2,000 draws takes 80% of the sites — without replacement — from every CHR × AF_bin stratum (genome-wide: every AF_bin stratum), and ρ is recomputed on the pooled draw. The reported ρ is the full-data estimate; the bars show the 2.5/97.5 percentile interval of the subsampling distribution. A result is called significant when that interval excludes zero; the directional rate mean(ρ≤0) is reported for reference only (floored at 1/(B+1)) and is not a conventional p-value.
The same subsampling analysis split by AF_bin — equal-width bins of the Non-Finnish-European allele frequency (AF_0-0.2 … AF_0.8-1.0), the stratification variable used by the reference analysis. ρ is the full-data estimate per bin; the interval is the 2.5/97.5 percentile of the subsampling distribution, and a bin is significant when its interval excludes zero.
log2|Evo2 Δ| stratified by functional region (G) and coding consequence (H), with group mean ± SEM, reference lines (selected-set / genome-wide), and one-sample tests (BH-adjusted).