Claude Skills by HolobiomicsLab

Use when when you have multi-branch deep learning architecture predicting related but distinct peptide properties (charge state, isotope count, retention time) from raw mass spectrum, and you want to leverage auxiliary task gradients to improve primary task learning.

Use when after RAMClustR clustering and do.findmain molecular weight inference are complete, when you need to submit the same inferred spectra to multiple third-party annotation tools (MSFinder and Sirius) that each require distinct file formats and cannot share a common intermediate representation.

Use when you have paired mass spectra and molecular structure datasets and need to train a model that jointly understands both modalities for tasks like structure elucidation. Specifically, use it when: (1) you have large unlabeled or weakly-labeled pretraining data with both spectra and molecules;

Use when you have a preprocessed peak table (feature matrix: samples × peaks) with known class labels or phenotype groupings, and you want to identify which individual peaks contribute most to classification or discrimination between groups.

Use when you have integrated, normalized lipidomic and metabolomic feature tables from the Multi-ABLE method or similar concurrent multiomics workflows, with matched sample phenotypes (e.

Use when immediately after parsing mzML files into (m/z, scan_number, intensity) tuples when you need to build mass tracks from raw MS1 spectra. Use it when working with high-resolution instruments (e.

Use when when building or extending a mass spectrometry data parser that must support multiple mzML storage formats (plain .mzML, indexed .mzML.gz, standard-compressed .mzML.

Use when you have mass spectrometry raw data in mzML format (including compressed variants like mzML.

Use when you have a compressed mzML file (mzML.gz or indexed gzip format) and need to extract a single spectrum or a small subset of spectra by their known numeric identifiers, rather than iterating through the entire file sequentially.

Use when after running AirdPro's CLI conversion pipeline (run-cli.sh with -i and -o arguments) to confirm the vendor raw file conversion to mzML has succeeded.

Use when you have an indexed gzip–compressed mzML file (mzML.gz with internal index structure) and need to retrieve and work with individual spectra or chromatograms by integer index without decompressing the entire file.

Use when you have one or more mzML files (XML-based mass spectrometry data) and need to convert them into mzPeak format for downstream analysis, archival, or integration with tools that consume Parquet-based spectra.

Use when when you have mass spectrometry data stored in mzPeak format (ZIP archive containing Parquet files) and need to read spectrum metadata, chromatogram metadata, and signal data (m/z and intensity arrays) for analysis. Use this skill if your input is an .

Use when you have raw mass spectrometry data (vendor formats, mzML, or existing mzPeak files) and need to: (1) convert to mzPeak format for long-term storage and interoperability across languages and tools; (2) load mzPeak spectrum or chromatogram data into memory as structured tables for analysis;

Use when you have an mzPeak file (uncompressed ZIP archive containing Parquet files) and need to extract and work with spectrum metadata (scan descriptions, precursors, selected ions), spectrum signal data (profile or centroid m/z and intensity arrays), or chromatogram data.

Use when you have mass spectrometry run data (spectra, chromatograms, instrument metadata) that must be stored in or recovered from the mzPeak format, or when you need to validate that a mzPeak implementation correctly supports both read and write paths for round-trip fidelity.

Use when you have a chemical structure (as SMILES string or identifier) and need to discover matching records across specialized natural product databases.

Use when when you have a large spectral library (thousands to millions of spectra represented as high-dimensional vectors) and need to search unknown query spectra against it, particularly under open modification search scenarios where all possible mass shifts must be considered.

Use when after identifying statistically significant features and assigning them to structural clusters (isotopologue groups, adduct groups, cross-assay links) and correlation clusters via MamsiStructSearch.

Use when when you need to benchmark multiple encoder types (e.g., FFN vs. GNN) on the same predictive task and require evidence that performance differences reflect genuine architectural trade-offs rather than suboptimal tuning.

Use when you have pre-trained MSGO model weights (PFAS or lipid variant) and a set of experimental mass spectra (either from reference evaluation sets or real wastewater/environmental samples) in CSV format with m/z and intensity pairs, and you need to obtain candidate molecular structures ranked.

Use when when building an end-to-end deep learning model that must predict multiple correlated peptide properties (charge, isotope count, retention time) simultaneously from mass spectrometry data, and you need a principled way to merge learned representations from separate task-specific branches.

Use when after implementing a neural network component that will feed into a downstream architecture (e.g., a transformer).

Use when use this skill after khipu has assigned observed ions to grid positions (isotope and adduct combinations). Apply it when you have a connected subnetwork of feature ions that have been matched to known isotope and adduct patterns and need to estimate the neutral mass of the parent compound.

Use when when comparing pathway analysis methods on metabolomics peak data and you need evidence that one method is more robust than another to the noise and missing peaks that are prevalent in real metabolomics experiments.

Use when when you have raw LC-HRMS profile-mode data (rt × mz intensity matrices) and need to detect chromatographic peaks using gradient-descent or local-maxima algorithms.

Use when you have high-dimensional replicate experiment data (e.

Use when implementing a backendInitialize() method for a custom MsBackend class and you have raw m/z and intensity vectors (from files, data frames, or lists) that must be converted into the internal NumericList representation required by the Spectra package.

Use when when extending the MsBackend virtual class to create a custom backend for mass spectrometry data, and you need to store peak m/z and intensity values as variable-length numeric vectors indexed by spectrum.

Use when you have loaded a numeric column (e.g., H/C ratio, O/C ratio, m/z value, or intensity) from a CSV file into Punc'data and need to render a histogram with appropriate bar spacing. The skill is triggered when the range of the column is small enough that default bin widths (1.

Use when when you need to create a plotting or visualization framework that must support multiple plot kinds (spectrum, chromatogram, mobilogram, peakmap) each backed by multiple rendering engines (matplotlib, bokeh, plotly), and you want to avoid combinatorial explosion of concrete classes while.

Use when query mass spectra do not confidently match unmodified peptides in the spectral library, or when you suspect the sample contains unknown or unexpected post-translational modifications.

Use when your query spectra contain peptides with unknown modifications (e.g., oxidation, phosphorylation, acetylation, or non-enzymatic modifications not specified in the search parameters), and you have a reference spectral library in mzML or mzXML format.

Use when when you have a query mass spectrum with unknown or unanticipated post-translational modifications and need to match it against an unmodified spectral library.

Use when when you need to make OpenMS C++ classes, functions, or data structures callable from Python code, or when verifying that a newly bound C++ component can be imported and instantiated without errors in a Python environment.

Use when when building a multi-backend visualization library where users specify both a plot type (spectrum, chromatogram, peakmap) and a backend (matplotlib for static output, Bokeh or Plotly for interactive), and you need to avoid code duplication across backends while keeping the user-facing API.

Use when a Python package has been relocated to a new repository location, reorganized to conform to new organizational standards (e.g., metabolomics-cloud conventions), or its dependencies, metadata, or CI/CD workflows have been modified.

Use when you need to verify that a Python package (or similar installable software) passes its declared integration test suite as a prerequisite to trusting its reliability in production or downstream analysis. Specifically, apply it when you observe a periodic testing CI workflow badge (e.

Use when when a package README or publication claims to install a large, fixed number of tools (e.g., 'approximately 89 tools') but does not enumerate them explicitly.

Use when after peak m/z values and molecular formulas have been extracted and pre-processed from FT-ICR MS data, and you need to reconstruct biochemical transformation networks ab initio.

Use when you have mass-spectrometry data in a Pandas DataFrame and need to expose plot kinds (spectrum, chromatogram, mobilogram, peakmap) as a `.plot(kind='...

Use when when you have raw mass spectrometry data (from mzML, Bruker .d, or CSV format) loaded into a Pandas DataFrame and need to ensure it has the correct column structure (m/z, retention time, intensity) before invoking pyOpenMS-Viz plotting functions like .plot(kind='spectrum'), .

Use when you have mass spectrometry data (retention time, m/z, intensity, or mobility dimensions) already loaded into a Pandas DataFrame and need to produce publication-ready or exploratory visualizations. Use this when you want to leverage Pandas' native .

Use when when you have a Spectra object backed by an on-disk MS data source (e.g., MsBackendMzR reading mzML, mzXML, or CDF files) and need to process large numbers of spectra in parallel or serial chunks.

Use when when a post-processing step (such as dereplication) risks removing true positive signals due to overly strict thresholds, and you need the option to retain raw or unfiltered output without code branching.

Use when setting up a LipidMatch analysis run and you need to select among three mutually-exclusive analysis modes (PFAS, Lipid, or Tween-positive detection). The trigger is application-level metadata or user specification that indicates which detection mode should execute.

Use when you have at least 3 raw mass spectrometry samples in open formats (mzML, mzXML, CDF) from untargeted metabolomics experiments and need to configure parameters for XCMS, MZmine2, or similar processing software.

Use when when implementing or extending a DSL parser (lexer + recursive descent or LALR parser) that accepts user-authored query strings.

Use when after marker identification or metabolite annotation has produced a curated list of compound IDs (e.g., KEGG IDs or CAS numbers) and you need to determine which metabolic pathways are statistically overrepresented in your dataset.

Use when you have run a metabolomics experiment with incomplete coverage of a reference pathway database (e.g., 10–100% of database metabolites detected), and you plan to use ORA for pathway enrichment. The skill is essential if your detection method has known sensitivity limits (e.