Nextflow Template

To run the workflow to test simply do

make run

To learn NextFlow checkout this documentation:

https://www.nextflow.io/docs/latest/index.html

Library Annotation (offline mode)

Library matches are annotated by bin/scripts/getGNPS_library_annotations.py. The forceoffline parameter (default Yes) controls how the compound metadata (Compound_Name, Smiles, Adduct, …) is resolved for each hit:

• forceoffline = No — annotate GNPS (CCMSLIB...) hits by calling the live GNPS API per hit.
• forceoffline = Yes — skip the API and resolve every hit locally from a library_summary.tsv that is built from the library .mgf files by bin/scripts/library_summary.py. This is much faster (no per-hit network calls).

NP Classifier Annotation

The npclassifier parameter (default Yes) adds npclassifier_pathway, npclassifier_superclass and npclassifier_class columns to the library results, by classifying each hit's SMILES via https://npclassifier.gnps2.org. It runs in both online and offline mode — offline mode otherwise skips all structure enrichment, which is why these columns disappeared from default runs.

Results are cached per unique SMILES, so the cost is one API call per distinct structure among the hits, not one per hit. Set npclassifier = No to drop the columns and make the annotation step fully network-free when forceoffline = Yes.

These columns are carried onto the network nodes as library_npclassifier_pathway, library_npclassifier_superclass and library_npclassifier_class in network.graphml.

Important — non-unique SCANS in the current GNPS2 libraries. Many of the libraries served from https://library.gnps2.org reuse the same SCANS= value across spectra (SCANS is a per-source-file counter, not a global unique id — e.g. GNPS-MSMLS.mgf has 863 spectra but only 56 distinct SCANS). library_summary.py must therefore read with mgf.read(..., use_index=False) and key on the real unique id (SPECTRUMID). Reading with index_by_scans=True collapses all spectra that share a SCANS value down to one, dropping ~94% of the library from the summary, so offline lookups miss and every field comes back N/A.

Shared code / modules

The scripts under bin/scripts/ are mostly symlinks into the GNPS_sharedcode submodule. GNPS_sharedcode (github.com/mwang87/GNPS_sharedcode) is being deprecated in favor of the consolidated NextflowModules submodule (github.com/Wang-Bioinformatics-Lab/NextflowModules), which vendors a copy of the shared code plus the workflow-specific, fixed scripts. library_summary.py was removed from GNPS_sharedcode as part of that deprecation, so it is checked in here as a real file (bin/scripts/library_summary.py) rather than a symlink. It is kept byte-for-byte identical to the canonical version at NextflowModules/bin/library_search/library_summary.py.

TODO: migrate this workflow onto the NextflowModules submodule (as LibrarySearch_Workflow already has) and repoint the remaining bin/scripts/ symlinks to it, so shared code has a single source of truth.

Installation

You will need to have conda, mamba, and nextflow installed.

Deployment to GNPS2

In order to deploy, we have a set of deployment tools that will enable deployment to the various gnps systems. To run the deployment, use the following commands from the deploy_gnps2 folder.

make deploy-prod

You might need to checkout the module, do this by running

git submodule init
git submodule update