Adversarial review of PR #846 found that kb/communities/Sphingobium_Nitrososphaera_Phenanthrene_Carbon_SynCom.yaml marked the random-forest / PLS-PM driver statement as IN_VITRO evidence.
The pot endpoints are experimental, but the claim that increased absolute abundance of carbon-fixing and PHE-degrading microorganisms primarily drove those effects comes from a statistical / machine-learning analysis over the pot data.
Fix: mark those EvidenceItems as COMPUTATIONAL and add computational_provenance for the RF and PLS-PM analysis.
Adversarial review of PR #846 found that kb/communities/Sphingobium_Nitrososphaera_Phenanthrene_Carbon_SynCom.yaml marked the random-forest / PLS-PM driver statement as IN_VITRO evidence.
The pot endpoints are experimental, but the claim that increased absolute abundance of carbon-fixing and PHE-degrading microorganisms primarily drove those effects comes from a statistical / machine-learning analysis over the pot data.
Fix: mark those EvidenceItems as COMPUTATIONAL and add computational_provenance for the RF and PLS-PM analysis.