ELF-NET Data S7: Novel domain discovery results from TARA-Oceans dark proteome analysis

ELF-NET Data S7: Novel domain discovery results Novel domain HMM profiles (.hmm), novel domain count matrix (2,044 samples × 33,950 novel domains), per-sample hmmsearch tblout from novel HMMs (.tar.gz), characterization table (prevalence, geographic breadth, nr-dark fraction), environmental correlation results (Spearman ρ, FDR-corrected q-values for 18 environmental variables), and novel-vs-Pfam effect size comparison. v2 addendum (2026-04-09): DataS7_structure_predictions_20260409.tar.gz adds ESMFold-predicted structures for the ten highest-ranked novel domains (10 .pdb files; sequences truncated to 400 aa per API limit), per-domain and per-residue pLDDT confidence scores, and Foldseek structural homology search results against the Protein Data Bank and the AlphaFold Database (best-hit TM-scores, E-values, and alignment statistics). Related to Figure 7, Figure S6, and STAR Methods. Restricted access during peer review; public upon publication.