HolobiomicsLab Feature-Based Molecular Networking with Annotation Propagation
Summary
MS2 in, a network-propagated annotation table out: molecular networking, seed annotation, chemical-class assignment, and topology-driven propagation of annotations and classes across molecular families.
When to use
Use when you have untargeted LC-MS/MS MS2 data and want to spread a handful of confident annotations across whole molecular families — build a feature-based molecular network, seed it with spectral-library and SIRIUS/CANOPUS annotations, then propagate compound classes and analogue annotations across network components (MolNetEnhancer / network annotation propagation) so unannotated nodes inherit chemically-plausible identities.
When NOT to use
- The data is not LC-MS.
- You need a single atomic step, not the full pipeline (use the leaf skill directly via the router).
Stages
Stage 1 — preprocess
Goal: raw mzML -> aligned feature table + MS2 exports (GNPS-FBMN mgf + SIRIUS mgf)
EDAM operation: operation_3215
Inputs: mzML · Outputs: feature-table, mgf/gnps-fbmn, mgf/sirius
Candidate leaf skills: peak-detection-and-mass-alignment (primary), cross-sample-feature-alignment, lcms-peak-detection-and-alignment, spectral-feature-table-generation, mass-spectrometry-feature-detection-validation
Tools (primary): MZmine2, Optimus, OpenMS
Other candidate tools: ISFrag, R, XCMS, CAMERA, JPA, MS-Convert, mzRAPP, MZmine 2, enviPat, Skyline, R (with mzRAPP library)
Grounding: 4 KB(s); DOIs: 10.1021/acs.analchem.1c01644, 10.1021/acs.jnatprod.7b00737, 10.1093/bioinformatics/btab231/6214530, 10.3390/metabo12030212
Stage 2 — network
Goal: MS2 spectra -> molecular family graph (modified cosine; GNPS-style components)
EDAM operation: operation_3214
Inputs: mgf/gnps-fbmn · Outputs: graphml, tsv
Candidate leaf skills: spectral-similarity-network-generation (primary), molecular-family-graph-construction, metabolomic-spectral-annotation-and-molecular-family-clustering, molecular-family-grouping-analysis, metabolomic-molecular-family-networking-gnps
Tools (primary): MZmine2, Optimus, GNPS, Cytoscape
Other candidate tools: nplinker, Python, pytest, antiSMASH, BiG-SCAPE, MIBiG, MS2LDA, PALS (Pathway Activity Level Scoring), GNPS (Global Natural Products Social Molecular Networking), MS2LDA (Mass2Motif Latent Dirichlet Allocation), PALS Viewer, conda, pip, BigScape
Grounding: 6 KB(s); DOIs: 10.1021/acs.jnatprod.7b00737, 10.1101/2024.10.11.617756, 10.1186/1471-2105-6-225, 10.1186/s40168-022-01444-3 …
Stage 3 — seed_annotate
Goal: MS2 spectra -> seed spectral-library annotations to propagate from
EDAM operation: operation_3631
Inputs: mgf/gnps-fbmn · Outputs: tsv
Candidate leaf skills: spectral-library-matching-annotation (primary), spectral-library-matching, spectral-library-molecular-networking, mass-spectrometry-library-ranking
Tools (primary): MSThunder, Windows, GNPS, MSConvert
Other candidate tools: microbeMASST, metadataMASST, plantMASST, tissueMASST, microbiomeMASST, foodMASST, GNPS_MASST, GNPS libraries, Fast Search API, MZmine, MASSBANK, DrugBANK, meRgeION2, RChemMass, MS2Compound, CFM-id, mssearchr, R, NIST API, MSHub, Python, Anaconda, Git, MSBERT, PyTorch, matchms, Spec2Vec
Grounding: 7 KB(s); DOIs: 10.1016/j.enceco.2025.07.022, 10.1021/acs.analchem.2c04343, 10.1021/acs.analchem.4c02426, 10.1021/jasms.5c00322 …
Stage 4 — class_annotate
Goal: MS2 spectra -> molecular formula + chemical class (SIRIUS / CANOPUS / NPClassifier)
EDAM operation: operation_3860
Inputs: mgf/sirius · Outputs: tsv
Candidate leaf skills: chemical-ontology-mapping (primary), spectral-feature-chemical-assignment, consensus-classification-reconciliation, structural-annotation-integration, chemical-classification-scheme-validation
Tools (primary): SIRIUS, NPClassifier, GNPS, CANOPUS, ClassyFire, ConCISE
Other candidate tools: Fiehn Labs ClassyFire Batch
Grounding: 1 KB(s); DOIs: 10.3390/metabo12121275
Stage 5 — propagate
Goal: spread seed annotations + chemical classes across molecular families
EDAM operation: operation_3434
Inputs: graphml, tsv, tsv · Outputs: tsv
Candidate leaf skills: graph-based-feature-annotation (primary), chemical-class-metadata-integration, molecular-network-node-annotation, molecular-network-attribute-enrichment, molecular-network-annotation-integration
Tools (primary): pyMolNetEnhancer, Python, RMolNetEnhancer, GNPS, MS2LDA, Cytoscape
Other candidate tools: ms2lda.org, MS2LDA (ms2lda.org)
Grounding: 1 KB(s); DOIs: 10.3390/metabo9070144
Stage 6 — consolidate
Goal: consolidate network family + seed + class + propagated annotations into one table
EDAM operation: operation_3434
Inputs: feature-table, graphml, tsv · Outputs: tsv
Candidate leaf skills: feature-metadata-annotation (primary), feature-consolidation-across-batches, untargeted-metabolomics-dataset-integration, lcms-feature-table-construction, feature-annotation-consolidation
Tools (primary): msFeaST, jupyter-notebook, msFeaST Dashboard bundle
Other candidate tools: R (>=), LargeMetabo, R, Matlab, M2S, Centwave, FeatureFinderMetabo, ADAP, ProteoWizard, MsFeatures, xcms, faahKO
Grounding: 5 KB(s); DOIs: 10.1021/ac051437y, 10.1021/acs.analchem.1c02687, 10.1021/acs.analchem.1c03592, 10.1093/bib/bbac455 …
Grounding
Each stage carries the kb_slugs/dois of the leaves it draws on. Ground any stage against its source paper with the collection's /ground command or bin/perspicacite_kb_bind.py (Perspicacité KB; serverless local-clone fallback).
Verification contract
workflow.yaml is gradable by asb solve-workflow (checkpoint mode). Each stage declares typed outputs; the final stage emits the master deliverable.
Provenance
Generated by compose_workflows.py (semantic binding + EDAM-aware primary selection). derived_from_workflows lists ASB per-paper workflows whose structure corroborated this pipeline — the eval-ablation set (SPEC §8). Staging only; promote via release_gate.py.