Ecotype Label Validation Benchmark

Why It Matters

Ecotype labels are useful only if they transfer beyond the conditions that produced them. The Atlas already records the tension: labels can support ecological synthesis, but they can also encode batch, host, or metadata artifacts.

Review Brief

What changed: ecotype assignments are being reused across host, plant, and environment pages, so validation needs to become a first-class benchmark rather than page-level caveat text.

Why review matters: reviewers should decide what validation tier is required before ecotypes support intervention, target, or broad ecology claims.

Evidence to inspect:

• Ecotype Assignments for reusable labels.
• Ecotype labels versus translational leakage for failure modes.
• ecotype_env_reanalysis, plant_microbiome_ecotypes, and ibd_phage_targeting for cross-domain reuse.

Questions for reviewers:

• Which holdout split matters most: study, host, geography, environment, or cohort?
• Should every ecotype label receive a validation tier?
• What leakage checks are mandatory before translational reuse?
• Which downstream page should be updated first if labels fail validation?

Evidence Base

This opportunity connects Ecotype Assignments, Environment Harmonization, Web of Microbes context, plant microbiome studies, and lab-field translation work. The benchmark should make the validation ladder visible rather than burying it in one notebook.

Work Package

Define standard splits by study, host, environment, geography, and metadata completeness. Re-evaluate label stability and predictive value under each split. Record which labels are exploratory, reusable within a domain, or ready for cross-domain interpretation.

Decision Use

If this benchmark succeeds, ecotype pages can cite concrete validation tiers. If labels fail, the Atlas should keep them as exploratory derived products and redirect downstream claims to narrower evidence.