{
  "created_utc": "2026-09-19T05:35:03.871877+00:00",
  "stage": "Independent native single-motif elimination extension; outcomes not yet inspected",
  "rationale": "After primary confirmation, the independently engineered loss-of-function set is large enough for a complementary test. If close motifs facilitate enhancement, removing a close motif should on average cause more loss.",
  "eligibility": "Only eliminateGATC_c2a variants with >=2 native consensus motifs, exactly one motif lost and no gained motifs, exact single-base change C to A. Exclude parents used in synthetic-addition analysis for an independent-parent check, while also report all parents. Report all distances and the subset with nearest remaining motif start >=10 bp to match addition-library spacing.",
  "analysis": "Outcome mutant minus original log2 activity, separately per backbone. Predictor nearest remaining motif start <=16 bp. Covariates native motif count (categorical), independent Table3 same-backbone Arabidopsis DOWN activity, original sequence GC and GC squared, source relative-position category, motif position/160 and identity of 4 flanking bases (C/G/T indicators at motif positions0,1,6,7). Parent-cluster standard errors. Minimum >=60 parents, >=20 close and >=40 distant losses for interpreting coefficients. Directional prediction negative; replication supported if both effects negative and both 95% CIs exclude zero. Failure will be reported and limits the scope to the synthetic-addition design; it cannot erase the held-out result or justify claiming physical cooperativity.",
  "gene_cluster_check": "Recompute primary and full-control uncertainty clustering by source gene (150 validation genes rather than153 parent fragments). Fixed effects remain parent-specific."
}
