CPR DPANN graph: qualify size-fraction attachment and CPR-only replication inference #1439

Episymbiotic_CPR_DPANN_Groundwater_Community.yaml, PMID:33495623.

Full primary main body and Methods read from the existing OA cache. The eight-site Northern California framework is not one defined three-member culture. Site differentiation does not isolate niche partitioning; genome recovery/coexistence is not colonization facilitation. Larger size fractions are an attachment proxy, not direct host identification. Cryo-TEM images show attached ultrasmall cells likely CPR/DPANN by size, not taxonomically resolved pairs or a selectively tested pili protein.

Higher iRep describes population genome-replication state in CPR; archaea were excluded because replication is not generally bidirectional. It is not a single-cell growth-rate or attachment-perturbation experiment. Keep the existing attachment-to-replication link HYPOTHESIZED and remove COMMENSALISM (host fitness untested). Preserve the three workflow-arrow removals recorded on 2026-10-02. Name canonical broad participants without assigning every lineage an effect.

Deduplication: 760 all-state issue titles/bodies and ignored/hidden-inclusive local review/research/history/tests/scripts/workflow files. Relevant #757, #382, #376, #369 and the existing CPR workflow-pruning history read. No exact duplicate in that scope; not every issue comment searched. Primary access limits are explicit above; no visual figures, separate supplements or raw data reviewed.