{
  "date": "2026-10-13",
  "study": "GSE289489 / EZbakR HEK293 SMG1 inhibition",
  "protocol": {
    "lockedAtUTC": "2026-10-13T15:58:23.132599+00:00",
    "models": {
      "M4_v5": {
        "path": "backend/models/m4-decay-response-v5/M4_v5.joblib",
        "sha256": "b565e5c5e1166764d44467442364e3f5011520f8f0438ae71eecb18142494266"
      },
      "ExtraTrees": {
        "path": "backend/models/m4-decay-response-v5/ExtraTrees.joblib",
        "sha256": "9b3b0d9ca94e41580b6927dfb03efaa3221ac5ba65c154f5e3ea3a3867a82acf"
      },
      "ridge": {
        "path": "backend/models/m4-decay-response-v5/ridge.joblib",
        "sha256": "1832ca7c0077b5c31291a7f5b7b36b428363ef6f466077dad169308a203dd9d4"
      }
    },
    "predictionSha256": "306c094bf21e022595be2780add8477111e3d7987f7f7df5fae61f751bb570e4",
    "geneMapSha256": "d7ad7fe5bbc3cc81a20f6053be92ef29d46c8bb9504a616d72cc5873bc0edbbc",
    "trainingSourceSha256": "08527f163c57d28fc248a78eb7f9c262dafed686300727a4257df2e90d2ff68b",
    "study": "GSE289489 / EZbakR HEK293 SMG1 inhibition",
    "endpoint": "log2(kdeg SMG1i / kdeg DMSO), gene-level GRAND-SLAM estimates processed by study authors with EZbakR",
    "sourceFile": "https://ndownloader.figshare.com/files/53217122",
    "sourceChoice": "Published Figure2.R explicitly uses EZbakRFit_GRANDSLAM_11j.rds for gene-level analysis. Chosen before numerical outcomes; avoids isoform-to-gene averaging. It is 21.7 MB versus the alternative 821 MB object, beyond available workspace capacity. Do not select an estimator by benchmark score.",
    "primaryMetric": "MAE; secondary Spearman and R2",
    "cohort": "All finite gene-level differences in the correct SMG1i/DMSO contrast with an exact stable Ensembl gene ID (version stripped) or unique exact reference gene-name mapping and frozen input predictions. Exclude ambiguous or duplicate mappings entirely. No significance, direction, magnitude or winning-subgroup filtering.",
    "targetTransform": "Published difference is natural-log kdeg contrast; divide by ln(2). Confirm reference and treatment from object metadata before scoring; stop if ambiguous.",
    "confidence": "2000 paired gene bootstrap resamples; baseline MAE minus M4 MAE; 97.5% intervals for two primary comparisons; descriptive conditional on published estimates.",
    "freshStudy": true,
    "directUPF1Holdout": false,
    "frozenBeforeOutcomeAccess": true,
    "limitations": [
      "Human SMG1 pathway transfer, not direct UPF1 depletion or therapeutic efficacy.",
      "Previously published aggregate findings were inspected for qualification; numerical gene outcomes were not opened before prediction lock.",
      "Public independent source study, no third-party blinded custody or raw-read duplicate audit.",
      "Existing input bank derives from development response availability and limits coverage; many genes may be previously seen.",
      "Uses published GRAND-SLAM/EZbakR gene estimates, not raw-read re-quantification; approximation and annotation differences remain."
    ]
  },
  "sourceRows": 19093,
  "unmapped": 2754,
  "ambiguousRowsExcluded": 0,
  "nonfiniteExcluded": 0,
  "evaluatedGenes": 11026,
  "targetGenesWithoutFeatures": 5313,
  "featureGenesWithoutTarget": 251,
  "seenDevelopmentGenes": 11026,
  "metrics": {
    "M4_v5": {
      "spearman": 0.4837737197133644,
      "mae": 0.30691217006207405,
      "r2": 0.06523335202724223
    },
    "ExtraTrees": {
      "spearman": 0.5971919435889725,
      "mae": 0.28264508225683,
      "r2": 0.22596433927698067
    },
    "ridge": {
      "spearman": 0.42651699337024584,
      "mae": 0.3216653456688388,
      "r2": -0.043831517131259545
    },
    "training_median": {
      "spearman": null,
      "mae": 0.30850374408515713,
      "r2": -0.0719493901606727
    }
  },
  "comparisons": [
    {
      "baseline": "ExtraTrees",
      "baselineMinusM4MAE": -0.02426708780524406,
      "adjustedPairedCi": [
        -0.02729512065613519,
        -0.021382157557990625
      ]
    },
    {
      "baseline": "ridge",
      "baselineMinusM4MAE": 0.014753175606764757,
      "adjustedPairedCi": [
        0.013064794289089913,
        0.016537115548691396
      ]
    }
  ],
  "superiorityEstablished": false,
  "directUPF1Superiority": false,
  "sourceSha256": "d370bb625358d8addc2394d5bc17da1e09e6285cb9c736ecaee8d50c222d7482",
  "m5Search": [
    {
      "id": "GSE304548",
      "url": "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE304548",
      "decision": "Human UPF1 knockdown iPSC RNA-seq; no matched RPF in this deposit."
    },
    {
      "id": "GSE304550",
      "url": "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE304550",
      "decision": "Human motor-neuron RNA-seq; associated RIP-seq measures RNA binding, not ribosome footprints. Not a qualifying TE holdout."
    },
    {
      "id": "GSE289489",
      "url": "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE289489",
      "decision": "Human metabolic-labeling decay and long-read RNA measurements; qualifies only for the separate M4 SMG1 decay test, not M5 translation efficiency."
    }
  ],
  "publicOnly": true,
  "m5Decision": "No new qualifying matched human UPF1 RNA/RPF cohort found. No new M5 accuracy claim or retraining on these incompatible endpoints. Existing v6 and assay-sensitivity failures remain visible."
}