Provenance-separated coverage
Integrated public-database records and newly curated primary-literature evidence remain independently stored and jointly searchable.
Inspired by Diana Triformis
A multifunctionalpeptidedatabase
for moonlight peptide discovery
PepDiana integrates peptide identities from 26 public resources with newly curated paper-level experimental observations while preserving each data layer and its provenance. Its name reflects Diana's many lunar phases and the multifunctionality of moonlight peptides.
Integrated data
Database-derived records with source-level provenance
Peptide identities
82,224
26 public resources across 25 stored namespaces
Cited publications
12,264
Distinct PMID or title-year publication identities
Activity records
376,740
Normalized source-database activity observations
Primary literature data
Newly curated paper-level identities and observations
Peptide identities
3,375
Primary-only sequence and modification-aware identities
Source publications
3,682
Primary publications represented in the evidence layer
Experimental observations
55,908
Positive, negative, threshold, and qualitative records
Activity and moonlight landscape
Top-level activity classes
17
Distinct non-unknown Level 1 activity classes
Secondary activity annotations
735
Distinct non-unknown Level 2 activity labels
Moonlight peptides
14,868
Identities supported by at least two Level 1 activity classes
Provenance-separated architecture
One research interface, three explicit data layers
PepDiana supports joint discovery without collapsing database-derived annotations, primary-literature peptide identities, and sequence-unresolved evidence into one ambiguous record type.
Why PepDiana
Data quality and standards
Comprehensive schema
Sequence, structure, modification, activity, assay, target, condition, literature, and provenance share explicit fields.
Record-level traceability
Integrated activities retain source IDs and URLs; primary observations retain paper, mention, and source-location links.
Standards-aware representations
ProForma 2.0, HELM 2.0, and canonical/isomeric SMILES include validation coverage and readable failure states.
Versioned descriptors
Physicochemical descriptors are release-computed with locked tool versions, assumptions, input hashes, and quality flags.
Evidence preservation
Negative, threshold, complex, and sequence-unresolved observations remain queryable instead of being silently discarded.
Database SearchSearch and filter curated multifunctional peptide records by sequence, bioactivity, source, and annotation quality.Explore 
Moonlight PeptidesExplore peptides with multiple biological activities across therapeutic contexts and disease areas.Explore 
Structure AnalysisInspect structural fields, molecular descriptors, standard notations, and external analysis links from detail pages.Explore 
Bioactivity ExplorerUse activity co-occurrence, source matrices, and analytics drilldowns to discover functional patterns.Explore 
Curated DatasetsTrace source databases, confidence levels, literature evidence, and export-ready subsets.Explore 
Help & APILearn browsing workflows, exports, citations, API access, and upcoming AI-agent usage patterns.Explore Activity landscape
Evidence density at a glance
Source coverage
dramp36,168
dbaasp24,712
camp24,243
satpdb19,192
conoserver9,251
thpdb6,370
apd6,278
plantpepdb6,171
ahtpdb5,978
Curated provenance
Every record keeps source identifiers and links visible.
Assay-ready browsing
Activity classes, units, organisms, and methods stay close to the sequence.
Cross-database joins
Source multiplicity is surfaced as a practical quality signal.
ML-friendly shape
FASTA export and compact tables support downstream modeling workflows.