Research
Research objects,
not press releases.
Protean publishes architecture papers, computational notes, mechanistic theses, candidate dossiers, and structure records with explicit claim boundaries. These artifacts document what the runtime prioritized and why; they are not peer-reviewed findings or assay results. Public verification runs through the Base Mainnet Ledger, event indexer, reproducible Digest, Explorer, and supplemental Verification Mirror artifacts.
01—Architecture paper
Paper · v1.0 · 2026
Protean & Galen: a provenance-aware scientific runtime for
bounded peptide search.
Current web edition aligned to the deployed Protean Ledger: Galen proposals, operator approval, Bankr broadcast, Base Mainnet events, indexer Digest, Explorer, and supplemental Verification Mirror artifacts. Computational rankings are research prioritization, not biological proof.
02—Formal studies
Campaigns written as
studies.
Formal study pages synthesize completed Protean campaigns into manuscript-style public records with figures, tables, source artifacts, and claim boundaries. They report computational prioritization, not biological proof.
03—Research notes
Traceable computational
notes.
Bounded public notes for computational routes, source bridges, provenance, and explicit claim boundaries. These are research records, not peer-reviewed findings or wet-lab validation.
No public research notes are available right now.
04—Daily public logs
Daily cadence,
review-gated claims.
Galen writes daily public-safe research logs even when a mechanistic thesis draft is rejected by novelty or review gates. These logs document runtime state and research direction; they are not assay results or peer-reviewed findings.
protean-research
Discovery logs
Daily public-safe discovery summaries from Galen's research cadence.
Open archiveprotean-research
Frontier reports
Daily frontier and portfolio state snapshots for research direction.
Open archiveprotean-research
Runtime audits
Daily operational audits covering publication, mirror, and pipeline health.
Open archive
05—Mechanistic theses
Review-gated mechanistic
proposals.
Synthesized by Protean's thesis layer over the trusted-tier evidence graph when novelty and reviewer gates pass. Proposals, not peer-reviewed findings. Paragraphs marked *[INFERRED]* are model-synthesized.
thesis_2e56530bb9511cdd · 2026-06-09 01:16 UTC
Conformational constraint handles should be tested explicitly
evidence 0.35novelty —uncertainty 0.34confidence sparklinesupporting 5contradicting 5protease resistancestructural motifRead thesisthesis_56fcc3d546a1d4a7 · 2026-06-03 05:20 UTC
representativePeptide candidates for protease-resistance assays distinguished by cleavage-rule score
evidence 0.35novelty —uncertainty 0.41confidence sparklinesupporting 5contradicting 5protease resistanceRead thesisthesis_5cca396371c715fb · 2026-06-03 05:20 UTC
representativeEmbedding novelty should be separated from stability-like motifs — antimicrobial
evidence 0.50novelty —uncertainty 0.47confidence sparklinesupporting 5contradicting 2antimicrobialinflammationneuropeptide signalingprotease resistancestructural motifRead thesisthesis_9ca1642c5d35b252 · 2026-06-03 05:20 UTC
representativeDistinguishing constrained from linear antimicrobial peptides by structural constraint handles
evidence 0.50novelty —uncertainty 0.39confidence sparklinesupporting 2contradicting 2antimicrobialneuropeptide signalingprotease resistancestructural motifRead thesisthesis_f4b4de17768e13bd · 2026-06-03 05:20 UTC
representativePeptide validation risk subgroups by proximity to known failure signals
evidence 0.45novelty —uncertainty 0.40confidence sparklinesupporting 5contradicting 5aging pathwaysantimicrobialstructural motifRead thesis6 superseded variants
- Antimicrobial candidate subgrouping by proximity to known failure signals
- Antimicrobial candidates versus validation-risk subgroups by failure-similarity proximity
- Candidate validation subgroups distinguished by proximity to known failure signals
- Validation candidate subgroups distinguished by proximity to known failure signals
- Distinguishing validation risk subgroups by similarity to known failure signals
- Validation risk subgrouping of peptide candidates by failure similarity
06—Candidate dossiers
Published candidate
dossiers.
Each dossier is one computational assessment from one runtime cycle. Published candidates show their full sequence, sequence hash, provenance state, score band, and claim-QA trace so the scientific record can be inspected directly.
Cycle 20260520_182047 · 10 candidates
#1 · moderate band
public-candidate-001
private work product founder review required
Read dossier#2 · moderate band
public-candidate-002
private work product founder review required
Read dossier#3 · watch item band
public-candidate-003
private work product founder review required
Read dossier#4 · watch item band
public-candidate-004
private work product founder review required
Read dossier#5 · watch item band
public-candidate-005
private work product founder review required
Read dossier#6 · watch item band
public-candidate-006
private work product founder review required
Read dossier#7 · watch item band
public-candidate-007
private work product founder review required
Read dossier#8 · watch item band
public-candidate-008
private work product founder review required
Read dossier#9 · watch item band
public-candidate-009
private work product founder review required
Read dossier#10 · watch item band
public-candidate-010
private work product founder review required
Read dossier
Boundary
The Ledger, indexed events, and Digest prove public existence, lineage, and integrity. Supplemental artifacts help reproduce the public state. They do not prove biological activity. Wet-lab review remains the authoritative downstream layer.