{
  "created_utc": "2026-09-03 18:36:53 UTC",
  "analysis": "inclusive reliability-weighted population-centroid Europe-only PCA panel",
  "inputs": {
    "pca_file": "/media/davide/5f3be2d1-52b6-4624-8349-65bed45c5868/newfiles/final_imputed_master_hg38/data/italy_balanced_inclusive_pca_20260817/heldout_projection_sensitivity_20260816/balanced_europe/balanced_europe_reference_coordinates.csv",
    "pca_md5": "64078b66da56380f501b332881db5c47",
    "count_file": "/media/davide/5f3be2d1-52b6-4624-8349-65bed45c5868/newfiles/final_imputed_master_hg38/data/italy_balanced_inclusive_pca_20260817/heldout_projection_sensitivity_20260816/balanced_europe/balanced_europe_reference_population_counts.csv",
    "count_md5": "45dcee49dbb9c4aeb3ed7a8850b821c9",
    "anchor_file": "/media/davide/5f3be2d1-52b6-4624-8349-65bed45c5868/newfiles/final_imputed_master_hg38/data/italy_balanced_inclusive_pca_20260817/europe_pca_rgb_maps/europe_population_anchor_definitions.csv",
    "anchor_md5": "b50a8ef0deeaee224c72553999201eb7",
    "genotype_qc_summary_file": "/media/davide/5f3be2d1-52b6-4624-8349-65bed45c5868/newfiles/final_imputed_master_hg38/data/italy_balanced_inclusive_pca_20260817/data/pca_population_weighting/population_weighted_pca_summary.json",
    "genotype_qc_summary_md5": "47ef1101bf4c183ed66db037a1dd27b9",
    "western_balkan_geography_source": "Kovacevic et al. 2014; doi:10.1371/journal.pone.0105090; GEO GSE59032",
    "british_geography_source": "/home/davide/AADRv66/v66.2M.aadr.PUB.anno",
    "british_geography_source_md5": "a810fa31bc6633b75bc99374bbf3f6bd",
    "british_collection_reference": "1000 Genomes GBR panel; Coriell MGP00003"
  },
  "genotype_quality_control": {
    "rule": "exclude reference samples with inbreeding F above 0.20 or genotype missingness above 0.10 before LD pruning and PCA fitting",
    "excluded_reference_samples": ["HG00124.SG", "NA20526.SG"],
    "reason": "all observed audit markers were homozygous; inbreeding F was 1.0"
  },
  "population_weighting": {
    "rule": "individual weight 1/max(population N, 30); population total influence min(1, N/30)",
    "axis_fit": "PCA fitted to one mean standardized-genotype vector per population",
    "bootstrap_replicates": 30,
    "PC1_median_aligned_correlation": 0.9937037384,
    "PC2_median_aligned_correlation": 0.95676336026,
    "PC3_median_aligned_correlation": 0.73206575666
  },
  "reference_individual_count": 2485,
  "reference_population_count": 38,
  "mapped_individual_count": 2386,
  "mapped_anchor_count": 123,
  "excluded_from_geographic_mapping": [
    {
      "population": "CEU",
      "n_samples": 99,
      "coordinate_status": "excluded from map",
      "anchor_rule": "Utah CEU collection has no defensible single European locality"
    }
  ],
  "geographic_anchor_rule": "Use audited Italian, French, Western Balkan and British localities where available; otherwise use a disclosed population or regional centre. Localities do not become PCA-fitting units.",
  "sign_orientation": {
    "rule": "use stored PCA coordinates verbatim; no sign flips",
    "red": "PC1 as stored",
    "green": "PC2 as stored",
    "blue": "PC3 as stored"
  },
  "rgb_scaling": {
    "method": "piecewise-linear empirical percentile transform",
    "reference": "one equally weighted centroid per fitting population, excluding CEU",
    "plotting_positions": "(rank - 1) / (N - 1), with average ranks for ties",
    "tail_rule": "values beyond the reference minimum or maximum are clipped to 0 or 1",
    "reference_population_count": 37,
    "reference_file": "/media/davide/5f3be2d1-52b6-4624-8349-65bed45c5868/newfiles/final_imputed_master_hg38/data/italy_balanced_inclusive_pca_20260817/europe_pca_rgb_maps/europe_rgb_population_percentile_reference.csv"
  },
  "single_component_scaling": {
    "method": "same piecewise-linear empirical percentile transform used by the RGB channels",
    "reference": "one equally weighted centroid per fitting population, excluding CEU",
    "display_limits": [0, 1],
    "raw_values_preserved": true
  },
  "interpolation": {
    "method": "inverse-distance weighting applied separately to PC1 PC2 and PC3",
    "candidate_powers": [1.5, 2, 2.5],
    "chosen_power": 2.5,
    "locality_reliability_weighting": "IDW kernel multiplied by min(n_samples, 5) / 5",
    "full_reliability_sample_count": 5,
    "selection": "minimum mean leave-one-anchor-out standardized RMSE across PC1-PC3",
    "grid_resolution_km": 40,
    "maximum_nearest_anchor_distance_km": 600,
    "land_mask": "Natural Earth European countries plus Russia Turkey and Cyprus cropped to the declared map extent",
    "interpolated_grid_cells": 5705,
    "maximum_realized_nearest_anchor_distance_km": 599.8711384688
  },
  "similarity_check": {
    "spearman_geographic_vs_rgb_distance": 0.76242244618
  },
  "caveat": "RGB values visualize model-dependent genetic structure and are not ancestry proportions or mixture estimates."
}