GenoLeWM-FX decision package¶

Status: final decision package for issues #257 and #265.

The GenoLeWM-FX pivot is stopped at the feasibility gate. No FX teacher cache, residual JEPA model, Hugging Face sweep, locked benchmark, model artifact release, paper update, or public demo ships from this experiment.

Authoritative artifacts:

Decision¶

The parent epic takes the kill path. The checked-in feasibility report found a public 11,400-row TraitGym source slice suitable for a cheap source-only probe, but no reproducible 10k-50k ref/alt teacher-delta cache from AlphaGenome, Borzoi, Enformer, ChromBPNet, or another functional teacher.

Continuing would require private credentials, heavyweight teacher setup, new assay curation, or newly published teacher-cache artifacts before the actual GenoLeWM-FX hypothesis could be tested. That does not justify medium or expensive training jobs.

Demo Decision¶

No FX demo ships. A demo would imply an available FX model or teacher residual score that this experiment did not produce. Existing GenoLeWM demos remain manifest-backed execution examples only and should not be described as evidence of useful planning behavior.

Paper Decision¶

No FX positive-result paper section is added. If referenced in future paper-facing text, this path should be described as a killed feasibility probe: public labels were available, but the required teacher-delta target was not reproducible under the contract.

Follow-Up Trajectory¶

The #257 kill decision remains correct for the teacher-inference path. A separate follow-up trajectory can test whether public precomputed Borzoi-derived scores rescue the idea without running a teacher. The trajectory now has two explicit lanes:

the default executable lane uses TraitGym's row-aligned Borzoi_L2_L2.plus.all score artifact for the matched complex-trait slice;
the optional provenance lane joins against the large statgen/fipip Borzoi table only when that table is explicitly staged locally.

The follow-up must start with an alignment and artifact-receipt audit against TraitGym identities, labels, splits, and the row-aligned Borzoi score vector. It must stop quickly if there is not a reproducible, public, checksum-backed 10k-50k matched slice. If the optional full fipip table join is not run, reports must say that directly and must not claim exact fipip table overlap.

That follow-up is tracked in #266 and documented in GenoLeWM-FX Borzoi rescue plan.

The first follow-up gate now has a source-controlled outcome: GenoLeWM-FX Borzoi alignment and overlap report. It records a go decision for the compact TraitGym-native row-aligned Borzoi score path, with 11,400 usable rows and no #268 blockers. That is an alignment/cache-feasibility result only; it is not a model-quality claim and it does not claim exact fipip overlap because the full fipip table was not staged for that run.

The second follow-up gate also has a source-controlled outcome: GenoLeWM-FX Borzoi score cache report. It records a manifest-backed 11,400-row cache with the direct TraitGym-native Borzoi score, source metadata columns needed for the baseline gate, and explicit fipip exact-join status. That cache opens the #270 baseline/saturation gate; it still does not open a model-quality or training claim.

The third follow-up gate has a source-controlled outcome: GenoLeWM-FX Borzoi baseline and saturation report. It records that the strongest simple baseline is not saturated on the locked holdout split and that #271 may proceed. The reported baseline numbers are gate evidence only; they are not a trained GenoLeWM-FX model result and do not support public model-quality claims.

The residual-model gate now has a source-controlled outcome: GenoLeWM-FX Borzoi residual model report. It records a small positive mean lift over the strongest simple baseline, but the paired confidence intervals cross zero. That completes the implementation experiment for #271, but it does not support a positive locked-result claim for #272.

The follow-up trajectory is now complete: GenoLeWM-FX Borzoi final outcome report. It publishes no_positive_claim_fragile_lift as the final outcome. The precomputed-Borzoi path produced source-controlled overlap, cache, baseline, and residual artifacts, but it does not ship an FX demo, paper-positive result, or public model-quality claim.

Child Issue Resolution¶

Issue	Resolution
#258	Completed by the source-controlled contract.
#259	Completed as a kill decision at feasibility.
#260	Not planned after the kill gate; no teacher cache should be built now.
#261	Not planned after the kill gate; no residual model target is selected.
#262	Not planned after the kill gate; no Stage B or Stage C job should launch.
#263	Not planned because the locked benchmark depends on the killed path.
#264	Completed by publishing the kill report instead of model artifacts.
#265	Completed by this no-demo/no-paper-update decision package.