Claude Skills by HolobiomicsLab

Use when when preparing augmented training data for a Siamese rescore model that must learn to rank correct molecular formulas above incorrect ones;

Use when when training a Siamese architecture rescore model for MS/MS-based molecular formula prediction and you have an imbalanced training set with far fewer negative than positive spectrum pairs.

Use when when you have a pre-trained model and need to report stable, generalizable performance on a fixed training set with multiple held-out test splits. Specifically: when you have 10 (or n) random query/reference splits on the same dataset (e.

Use when when you have raw tandem MS metabolomics data in vendor formats (.

Use when you have executed multiple NPDtools database search pipelines (Dereplicator, VarQuest, Dereplicator+, or MetaMiner in different modes) on identical test spectra or RiPP sequence inputs and need to understand their relative sensitivity, specificity, and complementarity.

Use when you have paired microbiome (16S rRNA/metagenomic) and metabolome (LC-MS/MS or similar) count data and need to evaluate how well a predictive model (e.g., neural network, Elastic Net) generalizes across samples.

Use when you have an experimental mass spectrum (query) and a set of molecular candidate structures, and you need to rank the candidates by how well their predicted spectral features match the query spectrum.

Use when you have downloaded a GNPS archive from either GNPS1 (https://gnps.ucsd.edu) or GNPS2 (https://gnps2.org) and need to parse spectra (spectra.mgf), molecular family networks (molecular_families.tsv), spectral library annotations (annotations.

Use when after feature extraction or feature alignment when you have FeatureSet or Sample objects that must be exported as CSV files for sharing, archival, or downstream analysis. Specifically: (1) when exporting single-sample feature tables from find_feature() output;

Use when you have generated or curated a lipid spectral library (with precursor m/z, adduct information, charge states, retention times, and fragmentation patterns) and need to export it for use in either Excalibur-based DDA experiments on an Orbitrap mass spectrometer, or in Skyline for targeted.

Use when you have a CSV file containing molecule definitions (chemical formula, m/z, intensity, retention time, or other peak properties) and need to prepare it for SMITER's simulation workflow.

Use when a practitioner has pre-computed features from an external feature-finding procedure (e.g., vendor software, alternative open-source tools) and wishes to incorporate them into PFΔScreen's PFAS prioritization pipeline without re-detecting features from raw mzML data.

Use when when you have LC-MS/MS data in Mascot Generic Format (mgf) files and need to identify compounds against a curated custom database (e.g., prepared using CFM-id for a specific metabolite class or organism) rather than relying on in-built commercial spectral libraries alone.

Use when after signal drift correction (step 4) has computed per-feature D-Ratio values, and before normalization (step 7).

Use when after loading a specXplore session data object (saved .

Use when you have preprocessed and normalized ROI feature data extracted from mzXML or mzML mass spectrometry files and seek to increase feature representation and robustness before statistical modeling or machine learning.

Use when you have a conceptual MS/MS fragmentation strategy (e.

Use when you have untargeted metabolomics mass spectrometry data (MS2 spectra with m/z values and intensities) and an existing knowledge-driven metabolite network, and you need to enhance annotation accuracy and coverage by leveraging experimental similarity patterns rather than relying solely on.

Use when after completing data merging, cleanup, and batch correction steps in the FBMN-STATS pipeline, when you have a processed feature quantification table combined with sample metadata in memory (R data frame or Python pandas DataFrame) and need to preserve it for multivariate statistical.

Use when immediately after importing raw mass spectrometry data from supported formats (mzML, mzXML, msp, metabolomics-USI, MGF, JSON) and before performing spectral similarity comparisons or statistical analysis.

Use when you have deconvoluted or processed MS/MS spectra from SWATH-MS data that need to be (1) ingested into tools requiring open formats (e.g., spectral library matching, metabolite identification pipelines), (2) archived in public repositories, or (3) shared across different analysis platforms.

Use when you have raw or partially processed metabolomics data (mzML/mzXML format) from LC-MS or GC-MS runs and need to apply standardized feature detection, alignment, and intensity normalization as part of a reproducible workflow.

Use when after obtaining structural clusters from the MAMSI framework using different parameter configurations (e.

Use when building or auditing a multi-instrument MS data processing system that must route different chromatography modes (LC, GC), ion mobility, or imaging modalities (MALDI) to distinct processing workflows.

Use when after executing a MassQL query that returns a tabulated results DataFrame (e.g., MS1 or MS2 scan metadata, peak intensities, retention times), and you need to produce visual summaries suitable for publication, presentation, or exploratory analysis.

Use when you have MS/MS spectral library data currently stored in file-based formats (JSON, CSV, binary, MGF, MSP) and need to migrate to database-backed storage to support fast queries by metadata filters (precursor m/z, ion mode, retention time) and computed similarity scores against query.

Use when you have one or more MS/MS spectra (query spectra in mzML, mzXML, or MGF format) and need to identify unknown compounds by comparing them against curated spectral databases organized by biological domain (microbe, plant, tissue, microbiome, food).

Use when you are implementing a new MsBackend subclass and need to store spectra metadata (sample names, retention times, precursor m/z, etc.) separately from peak data (m/z and intensity pairs) while maintaining row-wise alignment.

Use when you have mass spectrometry data arriving through heterogeneous input formats (Task ID from GNPS, Universal Spectrum Identifiers, or Feature-Based Molecular Networking identifiers) and need to load, validate, and store them as a single standardized dataset object for interactive peak.

Use when when you have a raw GNPS or other spectral library dataset with inconsistent or incomplete instrument annotations, and you need to verify or reproduce reported dataset split counts (e.g., training/test compound ratios). Apply this skill when an instrument allowlist fix (e.

Use when when you have partitioned public MS/MS files from MassIVE using the ReDU File Selector into one or more filtered groups (G1–G6) and need to verify that each group's file count complies with computational constraints before submitting to GNPS molecular networking (3000 file limit) or.

Use when when preparing MS/MS spectra for deep learning model training on a specific instrument type (e.g., Orbitrap, Q-TOF), and you need to verify that configuration-driven filtering (e.g., adding 'ftms' to an instrument allowlist) produces training and test sets of the expected size (e.

Use when you have raw or processed LC-MS/MS data from DDA mode acquisitions and need to extract, annotate, and structure MS/MS spectra with purity labels (or quality indicators) to serve as input to the DNMS2Purifier customized model training workflow, or to prepare data for purification of.

Use when you have a virtual chemical mixture (MS1 peaks) and need to prototype a new DDA acquisition strategy before testing on real instrumentation. Use this skill when you want to compare how different parameter combinations (e.g., TopN=3 vs TopN=5, isolation_width=0.5 Da vs 1.

Use when you have real mzML LC-MS/MS data (e.g., from a Beer sample or HMDB reference set) and want to test whether a proposed TopN DDA strategy (or variant) can accurately reproduce the observed acquisition patterns, or you want to compare multiple acquisition controllers on the same chemical.

Use when when you have LC-MS/MS data collected in DDA mode and suspect that MS/MS spectra contain chimeric (multiply-charged or co-fragmented) ion signals that will degrade downstream spectral matching, library searching, or metabolite identification.

Use when when you have raw DDA mass spectrometry data (mzML, mzXML, or netCDF format) where precursor ions have been fragmented and you need to associate each fragment ion back to its parent precursor ion to generate coherent, precursor-specific fragmentation spectra for chemical annotation.

Use when you have centroided DDA mzML data and have already detected MS1 features (either via pyOpenMS or external feature finding), and need to extract MS2 fragment m/z values and diagnostic patterns (e.

Use when you have MS/MS spectra (centroided m/z and intensity pairs) and corresponding MS1 precursor masses but lack reference spectra or a priori formula information.

Use when you have annotated MS/MS spectra in MGF format and need to identify peptide sequences that may not exist in reference protein databases—such as in immunopeptidomics, metaproteomics, paleoproteomics, venomics, or monoclonal antibody assembly workflows.

Use when when you have unknown MS/MS spectra with observed precursor m/z values and want to infer the molecular formula and adduct type (e.g., [M+H]+, [M+Na]+, [M+K]+) in a de novo setting without access to spectral libraries.

Use when when you have high-resolution LC-MS/MS data for an unknown metabolite or small molecule, have computed or measured the molecular ion mass and fragmentation spectrum, and require de-novo structure generation because the compound is absent from spectral libraries or structure databases.

Use when you have a trained shallow decision tree on ChemEcho feature vectors (sparse, high-dimensional representations of tandem mass spectra peaks and neutral losses) and need to convert it into an interpretable, deployable query for a domain-specific language like MassQL.

Use when you have tandem mass spectra data and need to predict a discrete molecular property (e.g., presence/absence of a sulfo group) while maintaining full interpretability of the decision logic.

Use when you have preprocessed and normalized LC-MS metabolomics data from multiple disease groups (e.g., healthy, disease-A, disease-B) and need to identify which m/z features or their patterns discriminate between phenotypes.

Use when you have UPLC-HRMS data (ThermoFisher, Agilent, or MSConvert-compatible format) from a water sample, a precursor m/z and retention time of interest, and want to annotate an unknown compound by predicting its molecular formula, structure, and name using deep learning scoring rather than.

Use when when you have MS/MS spectra from GNPS or other libraries and need to apply a pre-trained FIDDLE model (TCN formula predictor or Siamese rescore architecture) without training from scratch. Use this skill before running inference on new samples or benchmarks.

Use when after training a Siamese neural network on MS/MS spectrum pairs, use this skill to quantify prediction performance on a disjoint test set (e.g., 3600+ spectra from 500 unseen compounds).

Use when you have a trained deep learning model and want to quantify prediction uncertainty for each input pair or decision point. Use this when you need to identify low-confidence predictions (high IQR) and filter them out to reduce error in specific score ranges (e.

Use when you have a pretrained deep learning model, a reserved test set with ground-truth annotations, and need to evaluate prediction quality or generate embeddings for downstream analysis. Typical triggers: benchmarking a new model against classical baselines (e.