SNP–ML Consensus Panels (Top-20/50/120/240 & consensus_panel_full) — HapMap CEU/YRI

This dataset releases consensus SNP panels derived from correlation-ranked (Pearson) selection and PCA selection frequency under leakage-free MCCV for HapMap Phase II (CEU vs YRI). Files include Top-20 (main-text), Top-50, Top-120, Top-240, and the consensus_panel_full table, plus a data dictionary.Methods (brief): Pearson |r| ranking; PCA-120 constructed via EVR-weighted loadings fit on training folds only; consensus score = mean(percentile |r|, percentile PCA selection frequency). See the associated manuscript for full details.Columns: rank, rsid, chr, pos, block (A/B), block_id (A1..A25/B1..B22), sign(r), |r|, |r|-percentile, PCA-selection-frequency, gene/region (optional).License: Data under CC BY 4.0.Citation: Please cite this dataset DOI (10.5281/zenodo.17271720) and the manuscript.