Claude Skills by HolobiomicsLab

Use when after performing retention-time-based grouping on LC-MS data (e.g., with a 10–20 second window) and observing that many feature groups contain features with similar retention times but dissimilar abundance patterns across samples or different extracted ion chromatogram peak shapes.

Use when you have experimental RT measurements from a source chromatographic method and need to predict RTs for the same molecules on a target chromatographic method, but lack a large calibration dataset (typical scenario: 10–100 molecules with ground truth RTs on both methods).

Use when after running XCMS-based alignment on LC-MS datasets with hundreds of samples or data acquisition periods longer than a week, when the assumption that all m/z bins in the same sample share a single warping function is likely to fail.

Use when when ingesting raw or vendor-format mass spectrometry data files of unknown or mixed acquisition modality, and you need to automatically determine whether the input originated from liquid chromatography (LC), gas chromatography (GC), ion mobility spectrometry (IMS), or MS imaging (e.

Use when you need to generate blank or background-only .mzML files for method validation, when you want to create synthetic negative controls with realistic instrumental noise but no analyte peaks, or when you need to simulate serum matrix background (e.

Use when after running tardisPeaks() with screening_mode=TRUE on centroided .mzML LC-MS data, when you need to visually inspect whether the 10 target compounds (internal standards and endogenous metabolites) were correctly detected within their expected m/z and retention time windows.

Use when after screening-mode validation of your m/z and retention time (RT) windows has confirmed that targets are visible in your data. Use it when you have centroided .mzML files, a validated target compound table with adjusted expected RT values (e.

Use when when you have processed LC-MS data through XCMS alignment but suspect misaligned features due to retention-time drift over long acquisition periods (>1 week) or large sample batches (hundreds of samples), or when peak-filling produces unexpectedly high coefficient-of-variation (CV > 0.

Use when after loading centroided .mzML LC-MS data and defining a target list (compound ID, name, m/z, RT, polarity) when you need to: (1) automatically locate and integrate peaks for known compounds across multiple runs; (2) generate per-target and per-run quantitative metrics;

Use when you have raw LC-MS data (mzML or vendor format) and need to discover and characterize all chromatographic features present, without prior knowledge of target analytes.

Use when you have raw mzML files from an FTICR-MS or other direct-injection MS instrument and need to identify discrete chromatographic peaks across the m/z and retention-time dimensions. Use this skill when you must isolate individual ion signals before applying calibration corrections (e.

Use when you have multiple high-resolution mzML files from LC-MS/MS experiments (e.

Use when you have loaded raw LC-MS or direct-injection FTICR-MS data (in mzML or netCDF format) into an XCMSnExp object and need to identify individual chromatographic peaks before feature grouping.

Use when after constructing baseline-corrected mass tracks (either composite across samples or per-sample) when you need to identify individual chromatographic peaks for feature extraction in LC-MS or GC-MS metabolomics workflows.

Use when when you have LC-MS/MS raw data (mzML or netCDF format) and need to isolate a specific compound's signal based on its known or suspected m/z value and retention time range.

Use when analyzing GC-MS data containing overlapping chromatographic peaks—a common scenario in untargeted metabolomics and environmental screening where sample complexity or chromatographic resolution limitations cause co-elution of structurally similar or temporally proximate compounds.

Use when when you have a feature table (CSV with m/z and retention time columns) and corresponding LC-MS raw data files (.mzXML or .mzML) from an untargeted metabolomics experiment, and need to filter out spurious peaks before downstream statistical or biological interpretation.

Use when when processing untargeted LC-MS metabolomics data with XCMS and need to identify low-quality peak integrations that may introduce noise or bias into subsequent compound identification and quantification.

Use when after retention-time-based and abundance-correlation-based feature grouping have produced composite feature groups, and you need to identify which features within a group actually represent different compounds.

Use when running targeted peak detection on LC-MS data acquired with multiple overlapping m/z scan windows and observing distorted or periodically discontinuous peak profiles in EIC plots.

Use when when you have a feature table from LC-MS peak detection (e.g., output from MassCube's nontargeted peak segmentation step) and need to assess which features have adequate chromatographic separation from coeluting or nearby peaks.

Use when after extracting retention times from top MS1 features detected in an LC-MS run, and when you need to evaluate whether a given gradient time range (e.g., 0–30 minutes) is being used efficiently to separate compounds. Apply this skill as the objective function in gradient optimization (e.

Use when you have raw spectra files from a liquid chromatography experiment (in-house or external database) and need to adapt a pretrained GNN-RT model to predict retention times for your molecules. Preprocessing is the mandatory first step before any model training or transfer learning can proceed.

Use when when you have raw GC–MS or LC–MS data (m/z vs retention time chromatography-mass spectrometry maps) and need to identify analyte signals and marker features without conventional peak picking;

Use when you have a large mass spectrometry dataset stored across multiple mzML, mzXML, or CDF files and need to perform operations (e.g., normalization, filtering, feature extraction) on the full dataset but memory constraints prevent loading all peak data simultaneously.

Use when you need to automate testing and quality checks on code changes—specifically when pull requests or commits are made to a repository and you want to verify that builds succeed, test suites pass, and code quality metrics meet project standards before merging.

Use when after generating cross-spectrum negative examples via precursor m/z windowing and before training a rescore model (e.g., Siamese architecture in FIDDLE v2.0.0).

Use when preparing training batches for a neural network classifier on LCMS peak data where class counts are unequal (e.g., more high-quality peaks than low-quality peaks).

Use when after identifying putative BGC-encoded precursor peptides from a genome assembly via MetaMiner's BGC identifier, when preparing a RiPP structure database for downstream spectral matching via Dereplicator.

Use when you have labeled training data (e.g., pqm_development with 500 peaks and 89 samples) and need to select which of multiple classification algorithms (e.g., AdaBoost, Random Forest, SVM) and their hyperparameters (e.

Use when after training or evaluating a classification model (e.g., a Siamese neural network for spectrum similarity prediction) and obtaining a prediction array and corresponding ground-truth label array.

Use when after fitting and optimizing a MB-PLS model on training data, apply this skill to the held-out test set (typically 10% of the original sample) to obtain unbiased performance estimates.

Use when you have vendor raw mass spectrometry data files (e.g., .raw, .d, .wiff2 formats) that must be converted to mzML or Aird format, and you need to automate the conversion in a batch workflow, Docker container, or non-interactive environment.

Use when after feature clustering and drift correction (Gravity and Blueshift outputs) are complete and you need to produce a final, validated feature table ready for storage, archival, or downstream analysis.

Use when after density-based clustering (e.g., DBSCAN) has been performed on a sparse pairwise distance matrix derived from MS/MS spectra nearest neighbor indexes.

Use when after mass-to-charge matching has produced a large table of candidate KEGG metabolites with multiple adduct assignments per feature. Use it when you observe that multiple candidate ions co-elute and share similar intensity profiles, suggesting they derive from the same parent metabolite;

Use when when you have executed multiple clustering tools on the same tandem-MS dataset and need to quantitatively compare their performance using normalized, comparable metrics rather than raw cluster assignments alone.

Use when you have applied two different clustering or dendrogram-flattening methods (e.g., constant-threshold vs. silhouette-score optimization) to the same feature set and need to assess whether the two solutions assign features to clusters consistently.

Use when you have raw tandem MS metabolomics data (in mzML or MGF format) and wish to compare the clustering performance of two or more MS clustering tools on the same dataset.

Use when you have LC-MS feature tables (with m/z and retention time columns) paired with raw .mzXML or .mzML files, and you need to automatically validate which detected features correspond to genuine chromatographic peaks versus false positives or noise.

Use when you have preprocessed MS/MS spectral data (normalized peak intensities and m/z values) in memory or on disk, a trained CNN model checkpoint available, and you need to generate molecular embedding vectors for matching against a reference database of known metabolites.

Use when when you need to detect and classify peaks in LC-MS regions of interest (ROIs) and simultaneously localize their boundaries for area integration.

Use when after normalization of a metabolomic feature matrix but before statistical testing, when you have both QC (technical replicate) and non-QC (study) samples and need to remove features with unstable or poorly reproducible signal patterns.

Use when analyzing complex GC-MS mixtures where two or more chemical compounds elute at similar or identical retention times, producing overlapping or merged peaks in the raw chromatogram.

Use when when you have two independent predictions of categorical outcomes (up/down/no-change variation signs) across multiple sample pairs and need to measure agreement beyond what would be expected by chance.

Use when when you have uploaded a pre-analytical data table containing sample metadata, processing delay annotations (pre- and post-centrifugation times), and paired NMR metabolomic measurements for a plasma or serum cohort, and you need to determine how processing delays impact metabolite.

Use when when you have acquired ion mobility–mass spectrometry data (drift time and m/z dimensions) and need to convert observed drift times into calibrated CCS values.

Use when you have LC-IMS-MS/MS data with drift_time measurements and need to convert raw drift times into calibrated CCS values for structural annotation.

Use when you have a set of molecular structures in SMILES format that require CCS prediction for metabolite annotation in untargeted mass spectrometry workflows.

Use when when you have LC-IM-MS/MS data with measured collision cross section (CCS) values and m/z assignments, and you need to disambiguate sterol isomers (particularly N-Me derived unsaturated sterols) by matching against a curated database of predicted CCS values and MS/MS fragmentation patterns.