[
  {
    "dataset": "BBBP",
    "task": "clf",
    "structure": {
      "n_molecules": 1975,
      "provenance": {
        "rows": 2050,
        "unparsed": 11,
        "multicomponent_raw": 105,
        "unique_canonical": 1975,
        "exact_duplicate_rows": 0,
        "respelled_duplicates": 64
      },
      "acyclic_pct": 4.8,
      "murcko": 1024,
      "generic": 693,
      "singleton_pct": 76.3
    },
    "arms": [
      {
        "arm": "Random",
        "grouping": "random",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 1975,
        "morgan_mean": 0.563,
        "morgan_sd": 0.01,
        "maccs_mean": 0.823,
        "maccs_sd": 0.007,
        "near_dup_pct": 11.2,
        "near_dup_sd": 1.1,
        "score_mean": 0.916,
        "score_sd": 0.007,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=own)",
        "grouping": "scaffold__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 1119,
        "morgan_mean": 0.45,
        "morgan_sd": 0.04,
        "maccs_mean": 0.758,
        "maccs_sd": 0.057,
        "near_dup_pct": 1.3,
        "near_dup_sd": 0.6,
        "score_mean": 0.852,
        "score_sd": 0.048,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=pooled)",
        "grouping": "scaffold__pooled",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 1025,
        "morgan_mean": 0.474,
        "morgan_sd": 0.034,
        "maccs_mean": 0.789,
        "maccs_sd": 0.044,
        "near_dup_pct": 1.8,
        "near_dup_sd": 0.7,
        "score_mean": 0.876,
        "score_sd": 0.029,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=Butina)",
        "grouping": "scaffold__butina",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 1084,
        "morgan_mean": 0.442,
        "morgan_sd": 0.024,
        "maccs_mean": 0.753,
        "maccs_sd": 0.04,
        "near_dup_pct": 1.2,
        "near_dup_sd": 0.7,
        "score_mean": 0.864,
        "score_sd": 0.04,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (pooled), DeepChem order",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 1025,
        "morgan_mean": 0.436,
        "morgan_sd": 0.0,
        "maccs_mean": 0.765,
        "maccs_sd": 0.0,
        "near_dup_pct": 1.7,
        "near_dup_sd": 0.0,
        "score_mean": 0.781,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (pooled), other tie-break",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first_tiesarbitrary",
        "n_seeds": 1,
        "n_groups": 1025,
        "morgan_mean": 0.414,
        "morgan_sd": 0.0,
        "maccs_mean": 0.752,
        "maccs_sd": 0.0,
        "near_dup_pct": 1.3,
        "near_dup_sd": 0.0,
        "score_mean": null,
        "score_sd": null,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (own), DeepChem order",
        "grouping": "scaffold__own",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 1119,
        "morgan_mean": 0.406,
        "morgan_sd": 0.0,
        "maccs_mean": 0.718,
        "maccs_sd": 0.0,
        "near_dup_pct": 1.7,
        "near_dup_sd": 0.0,
        "score_mean": 0.769,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Generic framework (acyclic=own)",
        "grouping": "generic__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 788,
        "morgan_mean": 0.439,
        "morgan_sd": 0.028,
        "maccs_mean": 0.754,
        "maccs_sd": 0.037,
        "near_dup_pct": 1.2,
        "near_dup_sd": 0.6,
        "score_mean": 0.878,
        "score_sd": 0.045,
        "metric": "AUC"
      },
      {
        "arm": "Hybrid",
        "grouping": "hybrid",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 1149,
        "morgan_mean": 0.471,
        "morgan_sd": 0.013,
        "maccs_mean": 0.785,
        "maccs_sd": 0.011,
        "near_dup_pct": 2.5,
        "near_dup_sd": 0.6,
        "score_mean": 0.837,
        "score_sd": 0.027,
        "metric": "AUC"
      },
      {
        "arm": "Butina cluster",
        "grouping": "butina",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 796,
        "morgan_mean": 0.389,
        "morgan_sd": 0.016,
        "maccs_mean": 0.755,
        "maccs_sd": 0.007,
        "near_dup_pct": 0.8,
        "near_dup_sd": 0.6,
        "score_mean": 0.859,
        "score_sd": 0.026,
        "metric": "AUC"
      }
    ]
  },
  {
    "dataset": "Lipophilicity",
    "task": "reg",
    "structure": {
      "n_molecules": 4200,
      "provenance": {
        "rows": 4200,
        "unparsed": 0,
        "multicomponent_raw": 1,
        "unique_canonical": 4200,
        "exact_duplicate_rows": 0,
        "respelled_duplicates": 0
      },
      "acyclic_pct": 0.1,
      "murcko": 2407,
      "generic": 1517,
      "singleton_pct": 77.1
    },
    "arms": [
      {
        "arm": "Random",
        "grouping": "random",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 4200,
        "morgan_mean": 0.631,
        "morgan_sd": 0.008,
        "maccs_mean": 0.872,
        "maccs_sd": 0.003,
        "near_dup_pct": 22.4,
        "near_dup_sd": 1.2,
        "score_mean": 0.702,
        "score_sd": 0.017,
        "metric": "Spearman"
      },
      {
        "arm": "Scaffold (acyclic=own)",
        "grouping": "scaffold__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 2412,
        "morgan_mean": 0.556,
        "morgan_sd": 0.004,
        "maccs_mean": 0.843,
        "maccs_sd": 0.006,
        "near_dup_pct": 7.4,
        "near_dup_sd": 1.0,
        "score_mean": 0.634,
        "score_sd": 0.024,
        "metric": "Spearman"
      },
      {
        "arm": "Scaffold (acyclic=pooled)",
        "grouping": "scaffold__pooled",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 2408,
        "morgan_mean": 0.557,
        "morgan_sd": 0.008,
        "maccs_mean": 0.842,
        "maccs_sd": 0.009,
        "near_dup_pct": 7.7,
        "near_dup_sd": 0.7,
        "score_mean": 0.622,
        "score_sd": 0.013,
        "metric": "Spearman"
      },
      {
        "arm": "Scaffold (acyclic=Butina)",
        "grouping": "scaffold__butina",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 2412,
        "morgan_mean": 0.556,
        "morgan_sd": 0.004,
        "maccs_mean": 0.843,
        "maccs_sd": 0.006,
        "near_dup_pct": 7.4,
        "near_dup_sd": 1.0,
        "score_mean": 0.634,
        "score_sd": 0.024,
        "metric": "Spearman"
      },
      {
        "arm": "Scaffold (pooled), DeepChem order",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 2408,
        "morgan_mean": 0.52,
        "morgan_sd": 0.0,
        "maccs_mean": 0.828,
        "maccs_sd": 0.0,
        "near_dup_pct": 9.2,
        "near_dup_sd": 0.0,
        "score_mean": 0.58,
        "score_sd": 0.0,
        "metric": "Spearman"
      },
      {
        "arm": "Scaffold (pooled), other tie-break",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first_tiesarbitrary",
        "n_seeds": 1,
        "n_groups": 2408,
        "morgan_mean": 0.524,
        "morgan_sd": 0.0,
        "maccs_mean": 0.829,
        "maccs_sd": 0.0,
        "near_dup_pct": 8.1,
        "near_dup_sd": 0.0,
        "score_mean": 0.564,
        "score_sd": 0.0,
        "metric": "Spearman"
      },
      {
        "arm": "Scaffold (own), DeepChem order",
        "grouping": "scaffold__own",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 2412,
        "morgan_mean": 0.522,
        "morgan_sd": 0.0,
        "maccs_mean": 0.828,
        "maccs_sd": 0.0,
        "near_dup_pct": 9.2,
        "near_dup_sd": 0.0,
        "score_mean": 0.596,
        "score_sd": 0.0,
        "metric": "Spearman"
      },
      {
        "arm": "Generic framework (acyclic=own)",
        "grouping": "generic__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 1522,
        "morgan_mean": 0.512,
        "morgan_sd": 0.02,
        "maccs_mean": 0.818,
        "maccs_sd": 0.014,
        "near_dup_pct": 5.0,
        "near_dup_sd": 1.2,
        "score_mean": 0.579,
        "score_sd": 0.026,
        "metric": "Spearman"
      },
      {
        "arm": "Hybrid",
        "grouping": "hybrid",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 2412,
        "morgan_mean": 0.556,
        "morgan_sd": 0.004,
        "maccs_mean": 0.843,
        "maccs_sd": 0.006,
        "near_dup_pct": 7.4,
        "near_dup_sd": 1.0,
        "score_mean": 0.634,
        "score_sd": 0.024,
        "metric": "Spearman"
      },
      {
        "arm": "Butina cluster",
        "grouping": "butina",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 1269,
        "morgan_mean": 0.384,
        "morgan_sd": 0.012,
        "maccs_mean": 0.782,
        "maccs_sd": 0.006,
        "near_dup_pct": 0.6,
        "near_dup_sd": 0.4,
        "score_mean": 0.505,
        "score_sd": 0.045,
        "metric": "Spearman"
      }
    ]
  },
  {
    "dataset": "Tox21",
    "task": "clf",
    "structure": {
      "n_molecules": 7823,
      "provenance": {
        "rows": 7831,
        "unparsed": 8,
        "multicomponent_raw": 244,
        "unique_canonical": 7823,
        "exact_duplicate_rows": 0,
        "respelled_duplicates": 0
      },
      "acyclic_pct": 22.7,
      "murcko": 2324,
      "generic": 1393,
      "singleton_pct": 76.2
    },
    "arms": [
      {
        "arm": "Random",
        "grouping": "random",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 7823,
        "morgan_mean": 0.575,
        "morgan_sd": 0.002,
        "maccs_mean": 0.843,
        "maccs_sd": 0.002,
        "near_dup_pct": 12.1,
        "near_dup_sd": 0.6,
        "score_mean": 0.81,
        "score_sd": 0.04,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=own)",
        "grouping": "scaffold__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 4099,
        "morgan_mean": 0.505,
        "morgan_sd": 0.005,
        "maccs_mean": 0.801,
        "maccs_sd": 0.006,
        "near_dup_pct": 7.9,
        "near_dup_sd": 0.6,
        "score_mean": 0.826,
        "score_sd": 0.081,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=pooled)",
        "grouping": "scaffold__pooled",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 2325,
        "morgan_mean": 0.413,
        "morgan_sd": 0.077,
        "maccs_mean": 0.719,
        "maccs_sd": 0.068,
        "near_dup_pct": 2.0,
        "near_dup_sd": 2.3,
        "score_mean": 0.641,
        "score_sd": 0.126,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=Butina)",
        "grouping": "scaffold__butina",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 2844,
        "morgan_mean": 0.468,
        "morgan_sd": 0.01,
        "maccs_mean": 0.787,
        "maccs_sd": 0.008,
        "near_dup_pct": 2.9,
        "near_dup_sd": 0.6,
        "score_mean": 0.776,
        "score_sd": 0.052,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (pooled), DeepChem order",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 2325,
        "morgan_mean": 0.398,
        "morgan_sd": 0.0,
        "maccs_mean": 0.734,
        "maccs_sd": 0.0,
        "near_dup_pct": 1.1,
        "near_dup_sd": 0.0,
        "score_mean": 0.746,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (pooled), other tie-break",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first_tiesarbitrary",
        "n_seeds": 1,
        "n_groups": 2325,
        "morgan_mean": 0.397,
        "morgan_sd": 0.0,
        "maccs_mean": 0.737,
        "maccs_sd": 0.0,
        "near_dup_pct": 1.3,
        "near_dup_sd": 0.0,
        "score_mean": 0.744,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (own), DeepChem order",
        "grouping": "scaffold__own",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 4099,
        "morgan_mean": 0.508,
        "morgan_sd": 0.0,
        "maccs_mean": 0.791,
        "maccs_sd": 0.0,
        "near_dup_pct": 9.6,
        "near_dup_sd": 0.0,
        "score_mean": 0.679,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Generic framework (acyclic=own)",
        "grouping": "generic__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 3168,
        "morgan_mean": 0.506,
        "morgan_sd": 0.017,
        "maccs_mean": 0.799,
        "maccs_sd": 0.01,
        "near_dup_pct": 8.3,
        "near_dup_sd": 1.4,
        "score_mean": 0.754,
        "score_sd": 0.134,
        "metric": "AUC"
      },
      {
        "arm": "Hybrid",
        "grouping": "hybrid",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 3186,
        "morgan_mean": 0.497,
        "morgan_sd": 0.016,
        "maccs_mean": 0.803,
        "maccs_sd": 0.01,
        "near_dup_pct": 3.6,
        "near_dup_sd": 1.3,
        "score_mean": 0.768,
        "score_sd": 0.052,
        "metric": "AUC"
      },
      {
        "arm": "Butina cluster",
        "grouping": "butina",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 2689,
        "morgan_mean": 0.46,
        "morgan_sd": 0.011,
        "maccs_mean": 0.792,
        "maccs_sd": 0.004,
        "near_dup_pct": 2.1,
        "near_dup_sd": 0.9,
        "score_mean": 0.808,
        "score_sd": 0.103,
        "metric": "AUC"
      }
    ]
  },
  {
    "dataset": "HIV",
    "task": "clf",
    "structure": {
      "n_molecules": 41120,
      "provenance": {
        "rows": 41127,
        "unparsed": 7,
        "multicomponent_raw": 3086,
        "unique_canonical": 41120,
        "exact_duplicate_rows": 0,
        "respelled_duplicates": 0
      },
      "acyclic_pct": 3.8,
      "murcko": 19081,
      "generic": 10708,
      "singleton_pct": 74.9
    },
    "arms": [
      {
        "arm": "Random",
        "grouping": "random",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 41120,
        "morgan_mean": 0.603,
        "morgan_sd": 0.004,
        "maccs_mean": 0.869,
        "maccs_sd": 0.001,
        "near_dup_pct": 13.8,
        "near_dup_sd": 0.3,
        "score_mean": 0.819,
        "score_sd": 0.011,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=own)",
        "grouping": "scaffold__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 20663,
        "morgan_mean": 0.533,
        "morgan_sd": 0.004,
        "maccs_mean": 0.84,
        "maccs_sd": 0.003,
        "near_dup_pct": 6.0,
        "near_dup_sd": 0.5,
        "score_mean": 0.797,
        "score_sd": 0.037,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (acyclic=pooled)",
        "grouping": "scaffold__pooled",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 19082,
        "morgan_mean": 0.533,
        "morgan_sd": 0.004,
        "maccs_mean": 0.841,
        "maccs_sd": 0.001,
        "near_dup_pct": 5.7,
        "near_dup_sd": 0.6,
        "score_mean": 0.807,
        "score_sd": 0.019,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (pooled), DeepChem order",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 19082,
        "morgan_mean": 0.437,
        "morgan_sd": 0.0,
        "maccs_mean": 0.792,
        "maccs_sd": 0.0,
        "near_dup_pct": 2.2,
        "near_dup_sd": 0.0,
        "score_mean": 0.779,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (pooled), other tie-break",
        "grouping": "scaffold__pooled",
        "ordering": "largest_first_tiesarbitrary",
        "n_seeds": 1,
        "n_groups": 19082,
        "morgan_mean": 0.485,
        "morgan_sd": 0.0,
        "maccs_mean": 0.812,
        "maccs_sd": 0.0,
        "near_dup_pct": 3.7,
        "near_dup_sd": 0.0,
        "score_mean": 0.757,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Scaffold (own), DeepChem order",
        "grouping": "scaffold__own",
        "ordering": "largest_first",
        "n_seeds": 1,
        "n_groups": 20663,
        "morgan_mean": 0.428,
        "morgan_sd": 0.0,
        "maccs_mean": 0.78,
        "maccs_sd": 0.0,
        "near_dup_pct": 1.9,
        "near_dup_sd": 0.0,
        "score_mean": 0.769,
        "score_sd": 0.0,
        "metric": "AUC"
      },
      {
        "arm": "Generic framework (acyclic=own)",
        "grouping": "generic__own",
        "ordering": "shuffled",
        "n_seeds": 5,
        "n_groups": 12450,
        "morgan_mean": 0.517,
        "morgan_sd": 0.006,
        "maccs_mean": 0.827,
        "maccs_sd": 0.013,
        "near_dup_pct": 5.1,
        "near_dup_sd": 0.8,
        "score_mean": 0.792,
        "score_sd": 0.034,
        "metric": "AUC"
      }
    ]
  }
]