Recording a non-candidate so it is not repeatedly re-triaged as a good one. Found while working #183; no change is proposed to the record.
Drought_Rhizosphere_Iron_Actinobacteria_Community (CommunityMech:000235) scored 4 cultivation cues in the keyword triage — grown in, growth chamber, incubat, mesocosm — putting it among the strongest remaining candidates. Reading the source shows every one of those cues belongs to a sub-experiment on a different plant.
The record is the field sorghum rhizosphere under drought: community_origin: NATURAL, community_category: RHIZOSPHERE, environment rhizosphere. Its primary experiment is
a large-scale field experiment in which drought-stressed and watered control sorghum plants were sampled weekly across 17 time points (TPs) in California's Central Valley
The growth-chamber text is a validation experiment on maize:
maize tom1 mutants were grown in the growth chamber under both drought and control conditions
and the "microboxes ... four weeks" passage is a third, separate assay.
So curating growth_media / cultivation_setup here would attach maize growth-chamber conditions to a sorghum field community. Every snippet would be verbatim, just validate and just validate-references would both pass, and the record would assert conditions that were never applied to it — the #529 failure mode exactly, and this time the wrong experiment is in the same paper, so even checking that the source studies the community would not catch it.
Conclusion: not enrichable. The community was sampled from a field trial, which #183 itself says legitimately has no growth medium.
Generalisable point: cue counting locates papers that contain cultivation language, not papers that cultivated this community. scripts/cultivation_spans.py says as much in its docstring and deliberately extracts nothing automatically; this is the case that shows why. A same-paper wrong-experiment trap is invisible to the member-naming heuristic, because the members are named — just not in the passage the conditions come from.
Recording a non-candidate so it is not repeatedly re-triaged as a good one. Found while working #183; no change is proposed to the record.
Drought_Rhizosphere_Iron_Actinobacteria_Community(CommunityMech:000235) scored 4 cultivation cues in the keyword triage —grown in,growth chamber,incubat,mesocosm— putting it among the strongest remaining candidates. Reading the source shows every one of those cues belongs to a sub-experiment on a different plant.The record is the field sorghum rhizosphere under drought:
community_origin: NATURAL,community_category: RHIZOSPHERE, environmentrhizosphere. Its primary experiment isThe growth-chamber text is a validation experiment on maize:
and the "microboxes ... four weeks" passage is a third, separate assay.
So curating
growth_media/cultivation_setuphere would attach maize growth-chamber conditions to a sorghum field community. Every snippet would be verbatim,just validateandjust validate-referenceswould both pass, and the record would assert conditions that were never applied to it — the #529 failure mode exactly, and this time the wrong experiment is in the same paper, so even checking that the source studies the community would not catch it.Conclusion: not enrichable. The community was sampled from a field trial, which #183 itself says legitimately has no growth medium.
Generalisable point: cue counting locates papers that contain cultivation language, not papers that cultivated this community.
scripts/cultivation_spans.pysays as much in its docstring and deliberately extracts nothing automatically; this is the case that shows why. A same-paper wrong-experiment trap is invisible to the member-naming heuristic, because the members are named — just not in the passage the conditions come from.