M³KG v1.0: An integrated knowledge resource for Mongolian medicinal materials, formulas, natural products and quantitative target evidence

Overview M³KG v1.0 is a versioned, evidence-resolved knowledge resource designed to integrate traditional Mongolian medicine knowledge with pharmacognostic origins, formula compositions, natural products, molecular targets, and quantitative experimental bioactivity evidence. The resource is organized around standardized Mongolian medicinal-material identifiers and comprises three interoperable knowledge layers: M³KG, M³KG-Formula, and M³KG-Herb. The core M³KG layer represents Mongolian medicinal materials together with traditional properties, functions, indications, and pharmacognostic origins. M³KG-Formula structures Mongolian medicine formulas and links formula ingredients to standardized M³KG medicinal-material entities. M³KG-Herb extends plant-derived medicinal materials through normalized botanical-source mappings to natural products and experimentally supported molecular-target evidence. The principal cross-module evidence path represented in the dataset is: Formula → medicinal material → botanical origin → natural product → quantitative bioactivity evidence → molecular target Each transition represents a distinct type of evidence. Traditional-knowledge assertions, pharmacognostic-source relationships, natural-product occurrence records, and compound–target bioactivity measurements are therefore maintained separately rather than being interpreted as equivalent evidence. Data contents The release contains structured tables describing Mongolian medicinal materials, traditional terms, pharmacognostic origins, Mongolian medicine formulas, formula ingredients, medicinal-material mappings, botanical-source mappings, natural products, molecular targets, quantitative bioactivities, and cross-module relationships. Stable entity identifiers are used to support joins among modules. Quantitative bioactivity records preserve, where available, the original activity type, activity relation, reported value, measurement unit, assay context, molecular-target identifier, and literature or database provenance. Supporting documentation includes a field-level data dictionary, provenance manifest, quality-control report, release-version information, and reuse/licensing information. Evidence model M³KG v1.0 follows an evidence-resolved data model. In particular: a medicinal material–botanical source relationship represents pharmacognostic or taxonomic evidence; a botanical source–natural product relationship represents reported compound occurrence or isolation evidence; a natural product–target relationship is represented through an associated experimental bioactivity record; traditional functions and indications are maintained as traditional-knowledge assertions and are not treated as experimentally validated molecular phenotypes; computationally inferred relationships are not merged with quantitative experimental bioactivity records unless explicitly identified as a separate evidence class. Consequently, network connectivity within the resource should be interpreted as an evidence-supported association structure rather than proof of a causal pharmacological mechanism. Data standardization Medicinal-material names, formula ingredients, botanical names, chemical entities, molecular targets, and quantitative activity records were standardized using reproducible processing workflows. Plant-derived medicinal materials were linked to botanical source records using normalized scientific names and explicit matching procedures. Natural products retain external chemical identifiers where available, including InChIKey and database identifiers. Molecular targets retain external identifiers such as UniProt accessions where available. Compatible concentration-based bioactivity measurements were additionally normalized to nanomolar units for descriptive data analysis while preserving the originally reported value and unit. Activity endpoints and heterogeneous assay systems were not assumed to be directly interchangeable. Quality control The release was subjected to automated database and relationship-level quality-control procedures, including integrity checking, foreign-key validation, duplicate detection, orphan-entity checking, internal count consistency, botanical-scope validation, formula parsing checks, unresolved ingredient mapping identification, and bioactivity metadata validation. Potentially ambiguous or unresolved records were retained with explicit mapping or QC status rather than being resolved by speculative assignment. Detailed QC definitions and release-level results are provided in QC_report.csv. Reproducibility and versioning This Zenodo record represents the frozen M³KG v1.0 release used for reproducible analyses and manuscript-level statistics. The public M³KG, M³KG-Herb, and M³KG-Formula web resources may continue to evolve after publication. Consequently, analyses intended to reproduce published results should use this versioned release rather than assuming that the continuously updated web interfaces contain identical records. Future changes affecting data files will be released as new dataset versions. Third-party data and provenance Parts of the natural-product and bioactivity integration originate from external resources, including NPASS. Third-party source identifiers, citations, versions, and transformation information are documented in provenance_manifest.csv. Third-party content is not automatically relicensed by M³KG. Users should consult the terms and licenses of the corresponding source resources before redistributing or reusing source-derived records.