Claude Skills by HolobiomicsLab

Use when when analyzing high-resolution mass spectrometry data from natural-abundance or labeled peptides where multiple charge states (+2, +3, +4, etc.) may be present within the same m/z isolation window (e.g., 880–890 m/z).

Use when you have peak-picked features with m/z, drift_time, retention_time, and intensity columns, and you need to identify monoisotopic peaks and their charge-state-specific isotopologue members (e.g., singly charged C13-substituted species).

Use when training neural networks on MS/MS spectra (or similar scientific data) where you need to preserve model states that improve validation performance.

Use when after training a tandem mass spectrometry embedding model (such as MSBERT) on a reference spectral library (e.g., GNPS), apply this skill to confirm that the high-dimensional embedding vectors project into interpretable chemical space.

Use when you have computed low-dimensional embeddings (e.g., t-SNE coordinates) or clusterings of mass spectra and need to validate that the learned representation space organizes molecules by chemical taxonomy.

Use when you have a GNPS molecular network (GML or GraphML format) and wish to annotate network nodes with chemical class labels to support metabolite family interpretation. Use it specifically when you need to overlay chemical classification schemes (e.

Use when you have generated a GNPS molecular network (classical or feature-based workflow) and possess chemical class annotations (from GNPS library matching, ClassyFire, or other structural classifiers) that you wish to propagate onto network nodes and edges to enable chemical family-level.

Use when you have a GNPS mass spectral molecular network (in graphml or cytoscape format) and want to enrich its nodes with chemical class information derived from GNPS public library spectral matches.

Use when you have structural annotations from in silico tools (SIRIUS, CANOPUS) and spectral library matches from GNPS, but need to (1) reconcile conflicting or incomplete ClassyFire ontology assignments, or (2) substitute NPClassifier taxonomy when ClassyFire ontology data is unavailable from GNPS.

Use when your LC-MS feature table is incomplete or has low chemical coverage because traditional peak extraction algorithms (e.g., standard XCMS workflows) systematically miss features at m/z and retention time positions corresponding to known suspect compounds in your database.

Use when when you have mass-to-charge (m/z) values from mass spectrometry imaging or other MS experiments and need to assign molecular formulae with high precision, especially in spatially-resolved metabolomics where traditional LC-MS annotation methods are insufficient.

Use when you have an unknown MS/MS spectrum with a measured precursor m/z and want to generate a list of plausible chemical formula–adduct pairs to rank or filter downstream.

Use when when you have a loaded metabolite database (e.g., hmdb_compounds.p pickle file) and need to constrain the chemical space to a specific instrumental m/z range (e.

Use when after running formula inference on a benchmark dataset with known formula and adduct ground truth (e.g., NPLIB1, NIST20, or CASMI 2022). Apply this skill when you need to quantify ranking performance, isolate the contribution of specific model features (e.

Use when you need to feed chemical formulas into a neural network-based formula scorer (such as MIST-CF) that must learn data-dependent representations of formula structure and composition.

Use when you have tandem mass spectra with known molecular structures (SMILES, InChI, or chemical formula) and aim to train or evaluate a formula-level spectrum predictor.

Use when you have collections of chemical formulae (e.g., from SIRIUS decomposition or subformula labeling) derived from MS/MS spectra and need to feed them into a transformer encoder.

Use when when running a ViMMS Environment simulation with save_eval flag enabled and you need to correlate fragmentation events in the output mzML file back to their originating chemical compounds for downstream evaluation, optimization, or validation of acquisition strategies.

Use when when you have a metabolomics peak list (m/z values with optional retention times) from LC-MS or GC-MS and want to filter a computationally expanded chemical library to only compounds whose calculated masses (accounting for ionization adducts) fall within a defined tolerance of observed.

Use when when aggregating MS/MS spectra from multiple public repositories (GNPS, MassBank, Mona) or in-house sources with inconsistent metadata naming conventions, missing or malformed adduct annotations, or incomplete chemical structure annotations (SMILES/InChI/InChIKey).

Use when when you have a GNPS molecular network (graphml or cytoscape format) and wish to annotate it with chemical class labels or MS2LDA-derived mass2motifs to highlight shared structural features or chemical families across spectral clusters.

Use when you have MS/MS spectra with high chemical noise (spurious ions arising from incomplete ionization, in-source fragmentation, or instrument artifacts) and you possess accurate molecular formula or SMILES structure and adduct information for the precursor.

Use when you have MS/MS spectra contaminated with chemical noise (spurious fragment ions that do not correspond to real chemical bonds or rearrangements), a known or predicted molecular formula or SMILES structure for the precursor, the adduct type (e.

Use when your analysis has produced both in silico structural predictions (e.

Use when you have LC-MS/MS data preprocessed with MZmine2 into an MGF file (containing MS1 and MS2 spectra) and a feature table (peak areas per sample), and you want to relate MS1 features to each other based on predicted molecular substructures and chemical properties rather than arbitrary.

Use when you have untargeted LC-MS/MS metabolomic data (MS1 features with MS2 fragmentation spectra) preprocessed by MZmine2 into a feature table (BIOM format) and MGF file, and you want to construct a chemically-informed hierarchy of features for alpha/beta-diversity analysis or to group features.

Use when after training a sequence-to-sequence decoder that reconstructs SMILES strings from fixed-size embeddings (e.g., from MS/MS spectra), use this skill to measure reconstruction fidelity on held-out test embeddings.

Use when you have an unknown compound's mass spectrum (m/z peaks and intensities in .

Use when you have a known compound structure with validated MS/MS spectrum and a structural analog (modified version) with its own MS/MS spectrum, and you need to assess whether a modification site prediction method correctly identifies which atoms were altered.

Use when you have acquired composite fragmentation spectra from DIA experiments (MS^E, AIF, or SWATH-MS) that have been deconvoluted by IDSL.CSA, and you need to export the resolved spectra in a format suitable for chemical structure identification against reference libraries or spectral databases.

Use when you have MS/MS spectra from public repositories (e.g., GNPS) that lack chemical structure annotations (InChI/SMILES), or have incomplete/inconsistent annotations.

Use when when you have a set of molecular structures (N-Me derived unsaturated sterol lipids or structurally similar organic molecules with C=C bonds) represented as SMILES or molecular geometry files, and you need to train or apply a machine-learning model to predict an instrument-dependent.

Use when when you have MS/MS spectra with known chemical structures (InChIKeys or SMILES) and want to validate whether a novel or existing spectral similarity scoring method actually reflects true chemical structural similarity.

Use when you have a mass spectral library (EI or MS2 format) loaded into R via read_lib() and possess either MOL files (from Lib2NIST export) or an SDF file containing the corresponding chemical structures, but the library entries lack SMILES fields or have incomplete structure information.

Use when when you have access to source code or algorithmic documentation of a metabolite generation pipeline (such as MAGMa's job subproject) and need to understand, validate, or reconstruct the transformations that convert a parent compound into enumerated metabolite candidates.

Use when when you have enumerated a large pool of candidate chemical subformulae for observed fragment peaks (m/z values) within a mass tolerance window and need to eliminate chemically invalid candidates before ranking.

Use when you have a chemical substrate and need to predict its biotransformation products using rule-based metabolism prediction. This applies when: (1) you possess a library of biotransformation rules extracted from a curated database (e.

Use when when your DDA-mode LC-MS/MS data exhibits chimeric spectra patterns that differ systematically from the reference training set used in DNMS2Purifier, or when you wish to optimize purification sensitivity/specificity for your particular instrument, ionization method, or sample matrix.

Use when you have acquired LC-MS/MS data in Data-Dependent Acquisition (DDA) mode for untargeted metabolomics and suspect contamination from chimeric (co-fragmented) MS/MS spectra.

Use when when you need to simulate LC-MS/MS data for fragmentation strategy development and do not have (or wish to augment) real experimental chromatograms. Specifically: (1) you have a list of known or sampled chemical compounds with molecular formulae;

Use when you have sqMass files containing pre-extracted transition group chromatograms from DIA-MS experiments and need to interactively visualize individual peptide precursor chromatograms, inspect peak quality via Q-values (typically 1% FDR cutoff), and overlay peak boundaries or apply on-the-fly.

Use when when you have loaded a TransitionGroup (extracted ion chromatogram or mobilogram from DIA-MS data) and need to identify precise peak boundaries and apex positions for feature extraction.

Use when you have multiple LC-MS samples in mzML or mzXML format with variable retention times, at least one designated as a quality control (QC) reference file, and you need to correct for RT drift before peak/feature detection.

Use when you have mass spectrometry data loaded as a Pandas DataFrame with retention time and intensity columns, and you need to visualize the overall or mass-trace-specific signal intensity distribution across the chromatographic separation.

Use when after chromatographic peak detection (e.g. centWave) has been performed on LC-MS data and you need to group features that likely originate from the same compound.

Use when when processing a cohort of centroided mzML LC-MS files with high sample-to-sample retention time and m/z drift, and you need reproducible alignment of detected peaks across all samples before gap-filling and feature consolidation.

Use when after feature detection when you have a feature table with m/z, retention time, and intensity columns, and you need to group features into empirical compounds (putative metabolites) that account for isotopologue patterns and multiple adduct forms arising from a single underlying analyte.

Use when when ingesting raw mass spectrometry data from multiple instrument vendors or file formats into a metabolomics processing pipeline, and you need to expose spectral and chromatographic metadata through a single, consistent interface regardless of the source format's internal structure.

Use when after peak detection has been completed and a feature table has been generated in MetCohort.

Use when you have centroided mzML files from LC- or GC-HRMS instruments (acquired in data-dependent mode with ddMS2) and need to systematically identify chromatographic peaks, measure their mass and retention time coordinates, and quantify their intensities before applying PFAS-specific.