Upstream Complement Modulator Sweep — Computational Analysis (comp-018)¶
Frozen analysis archived to
./etc/experiments/comp-018-upstream-complement-modulator-sweep/wiki-archive.md(336 lines). This wiki stub remains so cross-references resolve and the page stays discoverable. Computational analyses are write-once artifacts; the daemon does not need to re-read them on every sweep, so the long content lives next to the experiment that produced it atetc/experiments/comp-018-upstream-complement-modulator-sweep/.⚠️ Read this before propagating any finding from this page. > > The brief that produced this experiment was contaminated by a contrived user-framing example ("if it's in rosemary I'll grow rosemary"), which biased the headline-promotion of rosmarinic acid as the singular Tier-1 candidate. An independent brief-scrubbed verification re-run (comp-020) ran 2026-05-08 and produced a meaningfully different ranking: three tied tier-1 candidates instead of one, with Helicteres benzofuran lignans (CH50 9/40 µM, single-paper anchor PMC6273495) as the highest-potency single hit in the corpus — beating rosmarinic acid by 4-20× on a matched assay. comp-020 also […]
Where the analysis lives:
- Full archived analysis: ./etc/experiments/comp-018-upstream-complement-modulator-sweep/wiki-archive.md
- Experiment directory (inputs, scripts, outputs): ./etc/experiments/comp-018-upstream-complement-modulator-sweep/
- Phase 2 multilingual + Helicteres replication + C1-INH engineering literature: ./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/
- Computational experiments index: computational-experiments.md
Phase 2 — Multilingual + Helicteres Replication + C1-INH Engineering Literature (2026-05-17)¶
Phase 2 deliverables live in
./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/. The Phase 1 contamination warning above remains in force; Phase 2 is additive, not a replacement.
Headline finding¶
Phase 2 confirms the comp-020 tier-1 ranking holds at the small-molecule level (rosmarinic acid + luteolin + Helicteres benzofuran lignans tied), surfaces a new orthogonal tier-1 candidate Phase 1 missed entirely — Houttuynia cordata polysaccharides (Chinese / Japanese / Korean / Vietnamese dietary herb, CH50 79-318 µg/mL across crude + purified fractions, multi-target C2 + C4 + C5, multiple in vivo precedents), and grounds the C1-INH parallel engineering thread on three load-bearing literature anchors:
- Bos 2003 (PMID 12758149, doi:10.1016/s1570-9639(03)00107-9) — Pichia pastoris yields 30-180 mg/L active rhC1-INH with same inhibitory capacity as plasma C1-INH despite different N-glycosylation. Direct precedent that yeast (Ascomycota) chassis is sufficient for active C1-INH production → koji thread plausible.
- Liu 2004 (PMID 15039314, PMC375168, doi:10.1128/IAI.72.4.1946-1955.2004) — N-deglycosylated C1-INH RETAINS protease-inhibitor function (C1s complex formation) but LOSES LPS-binding endotoxin protection. For OE platform CP0 use case (MSU-driven complement priming), protease-inhibitor function is load-bearing → major risk reduction for non-mammalian (koji, LBP) chassis with different or absent N-glycan.
- Cancian 2015 (PMID 26106828, doi:10.1097/ACI.0000000000000186) — Ruconest (conestat alfa, transgenic rabbit milk, non-human glycosylation) FDA-approved 2014 for HAE. Establishes regulatory precedent that FDA accepts recombinant C1-INH with non-human glycosylation provided protease-inhibitor function is demonstrated.
New tier-1 ranking (Phase 2 final)¶
| Rank | Compound class | Lead candidate | Potency | Dietary access | Anchor status |
|---|---|---|---|---|---|
| 1a | Phenolic acid | Rosmarinic acid | C3b 34 µM; CP 137-180 µM; bell-shape | FDA GRAS rosemary / lemon balm / spearmint | Multi-anchor |
| 1b | Flavone (multi-mechanism) | Luteolin | CH50 190 µM (CP), AP50 170 µM | Dietary (parsley, celery, honeysuckle) | Multi-anchor across XO + URAT1 + CP |
| 1c | Benzofuran sesquilignan glucoside | Helicteres compound 5 | CH50 9 µM (highest single-compound potency) | Not dietary; tropical Sterculiaceae | SINGLE-ANCHOR — replication INCONCLUSIVE per Phase 2 (./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/phase-2-helicteres-replication.json) |
| 1d (NEW; DUAL-CHOKEPOINT, upgraded 2026-05-19) | Pectic polysaccharide multi-target | Houttuynia cordata polysaccharide class (HCP / HCPM) | CH50 79-318 µg/mL across crude + purified (CP0); intestinal NLRP3↓ + tight junction↑ + TLR4-MD2 direct binding (CP1) | Widely dietary in Southeast Asia (鱼腥草 / どくだみ / diếp cá) | Multi-anchor: 3+ Chen Daofeng group papers (CP0) + Li 2025 PMC12254813 (dual CP0+CP1) + Yu 2026 PMC12937656 (TLR4 docking) + Cheng 2014 PMC7112369 (structure-dependent caveat) |
Helicteres replication outcome¶
INCONCLUSIVE / ANCHOR-STILL-SINGLE. Phase 2 found no independent group has reproduced the Yin 2016 (PMC6273495, doi:10.3390/molecules21111506) CH50 9/40 µM benzofuran lignan finding on a matched assay format. The Helicteres angustifolia genus has substantial cancer + antifibrotic + triterpene pharmacology literature but the upstream-complement angle is isolated to the single Yin 2016 paper. Structurally-adjacent benzofuran lignans (Styrax japonica egonol PMID 15643559, CP IC50 33 µM per Min 2004) are 3.7× weaker — possibly real exceptional pharmacology, possibly assay-format artifact. Independent wet-lab replication on a different laboratory's CH50 protocol remains the load-bearing risk-closure step. Full structured analysis at ./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/phase-2-helicteres-replication.json.
Multilingual-citation gap reframing¶
Phase 1 §10 listed "language barrier" as a limitation. Phase 2 reframes this: the upstream-complement natural product subfield is overwhelmingly Chinese-group (Chen Daofeng / Fudan, 14+ papers post-1999) and Japanese-group (Yamada / Kiyohara / Kitasato, 13+ papers since 1985) dominated, but these groups publish 80-95% in English-language journals (Planta Med, Carbohydr Res, Acta Pharm Sin B, Int J Biol Macromol, Phytomedicine). The actual barriers operational in Phase 1 were:
- Citation-network insularity — the Chen / Yamada matched-assay anti-complementary CH50/AP50 hemolytic discipline is a self-contained methodology stream that doesn't get heavily cited outside its immediate field
- Topic-discovery framing — traditional Chinese / Japanese terminology framings ("heat-clearing," "toxin-eliminating," 鱼腥草 vs Houttuynia cordata vs "anti-complementary polysaccharide") don't map cleanly to Western pharma mechanism-name queries
- Source-journal weighting — Zhongguo Zhong Yao Za Zhi and Acta Pharm Sin B are PubMed-indexed but underweighted in Web of Science impact-factor rankings
Implication for OE platform: the CLAUDE.md "global-multilingual research by default" rule operates correctly. The operational discipline is to query by traditional-formula-name + species-name + traditional-pathology-framing in addition to mechanism-name. A "C3 convertase inhibitor" query misses Houttuynia; a "Houttuynia cordata anti-complementary" query catches it. Mechanism-name is the wrong starting point for non-Western literature.
Phase 2 status¶
Phase 2 complete — 2026-05-17. Files:
- ./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/phase-2-multilingual-findings.md — Tier ½/3 candidates surfaced from non-English-corpus + Chinese-group + Japanese-group English-corpus
- ./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/phase-2-helicteres-replication.json — Helicteres benzofuran lignan replication attempt + verdict + recommendation
- ./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/phase-2-c1-inh-engineering-literature.md — C1-INH heterologous expression literature thread + top-3 anchors for comp-037
- ./etc/experiments/comp-018-upstream-complement-modulator-sweep/phase-2/phase-2-summary.md — synthesis + new tier-1 ranking + comp-037 inputs
Open follow-ups: Independent wet-lab Helicteres replication; Houttuynia in vitro replication on OE protocol; comp-037 commit with explicit reference to Bos 2003 + Liu 2004 + Cancian 2015; Phase 3 multilingual extension for other mechanism classes; DeepSeek V4-Pro translation cross-check for 4 Chinese-language Phase 2 sources flagged [TRANSLATION-SINGLE-MODEL].