--- profile: elgora_markdown_bounty_challenge_v0 escrow_amount: "1000000" submission_deadline: 1789614000 payout_policy: winner_take_all --- # Does the published BCL11A association evidence support transfer across sickle-cell populations? ## Summary Produce a reproducible aggregate-data audit of cross-population evidence for BCL11A as a sickle-cell research target. Determine which conclusions about direction, magnitude and transferability are supported by the fixed studies below, and which remain unresolved. A well-supported negative or inconclusive conclusion is eligible. ## Challenge details The OpenLabs project asks whether GWAS supports BCL11A as a target. This bounty purchases a focused scientific audit, not a new discovery, a clinical recommendation, or a reproduction using unavailable participant records. Analyze all six variants and all four study populations in Table 2 of Sebastiani et al. (2015). Use Bauer et al. (2013) to distinguish association evidence from functional evidence. Treat the two papers as a closed evidence set, not a systematic review of current therapeutics. Assess whether direction and statistical precision support transferability, whether reported effect magnitudes can validly be compared, and whether the evidence explains between-population HbF differences. Identify any transcription or denominator inconsistencies that prevent a stronger conclusion. Do not silently repair a source. ## What you need to submit (Deliverables) - `evidence.csv`: one row per variant and population from Sebastiani Table 2, recording source location, reported allele coding, beta, standard error, p value, response scale, genotype-count text and any ambiguity. Preserve the reported text beside any parsed value. A missing or ambiguous value must have a reason. - `analysis.py` or `analysis.R`: executable analysis of the submitted CSV, with its environment and invocation described in `README.md`. Compute approximate normal 95% intervals from beta and standard error, direction agreement across populations, and nominal versus Bonferroni significance across the 24 variant-population tests using the reported p values. This multiplicity analysis is a sensitivity analysis, not a replacement for the study's original hypothesis structure. Output `results.csv` containing these results. Do not pool coefficients measured on different response scales. - `audit.md`: a scientific argument answering the question, linking every material factual conclusion to paper section/table/figure and the relevant computed result. Include a source-discrepancy table; distinguish direction consistency, magnitude comparability, non-significant findings and evidence of equivalence; evaluate age, assay, population and dependence limitations; and explain what the functional paper adds and does not establish. State which additional evidence would resolve the most important remaining uncertainty, without claiming it was collected. - `README.md`: reproducibility instructions, dependencies, input provenance and any incomplete verification. All required files must be included in the Submission; links are citations, not replacements for files. ## Inputs, Materials and References Required fixed scientific inputs are the publicly accessible PMC article versions associated with these identifiers, as available on 17 September 2026. Their named tables, methods and figures govern this analysis; linked supplements are optional, and no participant-level data or author contact is required. 1. Sebastiani et al., *BCL11A Enhancer Haplotypes and Fetal Hemoglobin in Sickle Cell Anemia*, 2015, DOI 10.1016/j.bcmd.2015.01.001, PMCID PMC4341902: https://pmc.ncbi.nlm.nih.gov/articles/PMC4341902/ . Table 2 is the numerical input; Table 1, methods, results and discussion supply interpretation. Preserve inconsistencies between article representations rather than choosing an unsupported correction. 2. Bauer et al., *An erythroid enhancer of BCL11A subject to genetic variation determines fetal hemoglobin level*, 2013, DOI 10.1126/science.1242088, PMCID PMC4018826: https://pmc.ncbi.nlm.nih.gov/articles/PMC4018826/ . This supplies functional context only. Background project, not an additional acceptance source: https://openlabs-git-codex-openlabs-elgora-adapter-bio-xyz.vercel.app/projects/0108093d-9241-44da-8e95-61cd7dd64b05 . This bounty is an independent research contribution and does not claim project-owner endorsement. ## Acceptance Criteria 1. The evidence table covers the complete 24-row numerical scope. Reported beta, standard error and p values must faithfully match the fixed source; an ambiguous source string may remain unparsed if explicitly recorded and excluded only from the affected calculation. Invented corrections fail. Genotype-count inconsistencies do not excuse omission of otherwise readable beta/SE/p values. 2. Running the submitted analysis on the submitted table reproduces its results. Intervals use beta ± 1.96 SE on the reported response scale; significance uses p < 0.05 and p < 0.05/24 respectively. Missing values remain missing. Numerical outputs must agree with those definitions within 0.000001 absolute tolerance before display rounding. 3. The audit separates statistical direction from clinical effect and does not pool or directly equate incompatible scales. It explicitly considers that correlated tests, differing samples, measurement and transformations limit the sensitivity analysis and transferability claims. 4. The argument tests the cross-population claim against both supporting and conflicting evidence in the fixed sources. It must distinguish absence of statistical significance from proof of no effect or equivalence, and distinguish association, functional mechanism and clinical benefit. A citation list or generic summary without this adjudication fails. 5. The discrepancy table identifies material internal inconsistencies encountered in the required extraction, quotes only the brief relevant field text, gives exact locations and states their consequence for the analysis. The audit must not claim individual-level replication, new experimental validation, treatment safety or efficacy from this aggregate audit. ## How is the winner selected? Only Submissions meeting every acceptance criterion qualify. Compare eligible Submissions in this order: (1) fewer material factual or analytic errors, where material means capable of changing the transferability conclusion; (2) more complete handling of source inconsistencies and credible alternative explanations relevant to that conclusion; (3) stronger traceability from each conclusion to the exact source and computation. Additional length, unsupported sophistication and a positive conclusion earn no preference. If still tied, the earlier on-chain Submission wins; if its timestamp is also tied, the numerically lower Submission identifier wins. If only one qualifies, it wins; if none qualifies, the outcome is no_valid_submission. ## Out Of Scope New wet-lab work, participant-level inference, clinical advice, updated treatment recommendations, and claims to have established a new BCL11A therapy.