Claude Skills by HolobiomicsLab

Use when you have collected or parsed 1H and 13C NMR peak data (chemical shift values and intensities) and need to submit it to the SMART 3 /api/smart3/search endpoint or similar TensorFlow Serving-backed molecular classification system that expects peaks as JSON rather than raw spectroscopic files.

Use when when preparing an input spectral peak list for NMRformer or similar Transformer-based peak assignment models, and you have observed peaks whose chemical shifts do not align with known metabolite signatures.

Use when after peak clustering has been completed in pyINETA (i.

Use when you have preprocessed 1D NMR spectra (¹H and/or ¹³C) and need to extract spectral features for molecular structure inference on molecules with up to 19 heavy atoms. The skill is necessary as the first stage before fragment assembly or connectivity prediction;

Use when you have loaded raw NMR or MS metabolomic abundance data into a SummarizedExperiment object and need to assess feature reproducibility before downstream association modeling.

Use when you have per-feature CV values from quality control analysis of NMR or MS metabolomic data and need to: (1) establish whether your dataset meets FDA reproducibility standards for downstream biomarker discovery or quantification; (2) benchmark data quality against regulatory thresholds;

Use when you have NMR metabolite measurements from peripheral blood samples (plasma/serum) paired with processing delay metadata (pre-centrifugation and post-centrifugation times) and need to benchmark metabolic parameter stability across delay windows.

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 processing mwTab metabolomics data files with variable column naming conventions (e.

Use when when you have installed a Python package and need to verify that its command-line entry point is accessible, or when you need to discover available commands and options for NMR data processing workflows without consulting external documentation.

Use when you have an NMR mixture spectrum (1D or 2D) and a library of reference spectra for pure compounds, and you need to identify which compounds are present and in what proportions.

Use when after peak picking has identified individual signals in an INADEQUATE NMR spectrum, apply this skill when you need to collapse thousands of individual peaks into fewer, more interpretable compound-level peak networks.

Use when after training a FlavorFormer model end-to-end with weighted loss on 1H NMR spectra and compound labels, apply this skill to a held-out test set to measure compound identification accuracy and ranking quality.

Use when after filtering a peak table to remove mispicked ions, group contaminants, and low-replicability features, you have a curated feature list with m/z, retention time, and MS/MS spectra ready for annotation.

Use when after peak picking on INADEQUATE NMR spectra when you have a set of peak coordinates and intensities and need to determine which peaks belong to the same molecular compound.

Use when when designing a dataset storage layer that must handle variable dataset sizes, block layouts, and platform-specific constraints (e.g., Windows vs. non-Windows). Specifically: (1) you have parsed dataset metadata including cache-file flags, total point counts, and block configuration;

Use when when testing associations between metabolic features (from NMR or MS) and a phenotype of interest (e.g., BMI, disease status) in a cohort where age, gender, or clinical confounders are known to correlate with both the metabolite and phenotype.

Use when you have a Dockerfile and source repository for a bioinformatics tool (e.g., CloMet) and need to verify that the tool can be containerized, deployed, and made executable in an isolated environment.

Use when you have 1H NMR spectral tensors as input and need to extract local features (e.g., peak patterns, signal neighborhoods) before applying attention-based or sequence-level processing.

Use when after identifying statistically significant features within individual LC-MS assays (e.g., via MB-VIP and permutation testing), use this skill when you have multiple parallel assays acquired in complementary ionization modes (e.

Use when when you have per-feature quality metrics (such as CV values from NMR or MS reproducibility analysis) and need to: (1) confirm that a specified proportion of features meet regulatory thresholds (e.g., 99% < 0.30, 92% < 0.15 for CV);

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 you have raw spectroscopic datasets from heterogeneous sources (multiple Zenodo repositories with different file formats and scales) that must be jointly normalized, deduplicated, and aligned by molecular identifier to feed a multimodal deep learning architecture.

Use when you have NMR dataset metadata (cache-file flag, total point count, block layout configuration) and need to select an appropriate storage backend that balances memory efficiency, access patterns, and platform-specific constraints.

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 an unknown compound's mass spectrum (m/z peaks and intensities in .mgf or equivalent format with mandatory PRECURSOR_MZ and IONMODE tags) and need to identify structurally related metabolites from a reference database.

Use when when you need to document or reproduce a Python-based research application (or any package-dependent workflow) and discover that the original publication or repository specifies dependencies without versions, or when you want to verify that a documented set of pinned versions can be.

Use when when you have multi-modal spectroscopic data (IR, Raman, UV-Vis, mass spectra, or NMR) and need to recover the underlying molecular structure without relying on finite spectral libraries or autoregressive SMILES generation.

Use when you need to deploy CloMet for the first time on a new system, or when you want to ensure reproducible execution of metabolomics data harmonization tasks without manual dependency management.

Use when you have a formula-assigned FT-ICR MS dataset (molecular formulas already assigned to individual mass features) and seek to understand the chemodiversity landscape and transformation relationships within DOM.

Use when you have an NMR mixture spectrum and a library of single-compound reference spectra, and you need to identify which compounds are present in the mixture and their abundances.

Use when when reading a binary file format with a magic integer or fixed checksum field at a known offset, and endianness is not explicitly declared in file metadata or header comments.

Use when when you have obtained a Python package from a repository (e.g., via git clone) and need to confirm that the documented Python version constraint and pinned dependency versions are sufficient to execute the package's main entry point (typically main.py or a console script).

Use when you have completed NMR data quality control analysis and possess per-feature CV values, and you need to formally assess whether the metabolomic dataset meets FDA regulatory standards for downstream biomarker discovery or quantitative assays.

Use when you have detected feature tables from multiple LC-IMS-MS/MS samples and need to establish a unified feature catalog in which each row represents a distinct molecular entity observed across one or more samples, with harmonized m/z, drift time, and retention time coordinates.

Use when after simulating and convolving individual metabolite multiplets with realistic lineshapes (Lorentzian or Gaussian) and combining them into a single time-domain FID array.

Use when when you need to construct a POST endpoint that ingests raw spectral data files from multiple vendor formats (jcamp, RAW, mzML) and must standardize them for downstream processing.

Use when you have a raw NV (NMRViewJ) binary file and need to confirm it is well-formed before parsing or processing.

Use when when a web service must accept spectroscopic measurements (NMR peaks, IR/MS metadata) submitted as form-encoded POST parameters from a client, and those parameters need to be validated, reformatted into an external API query format, or passed to a downstream predictor service.

Use when when you have 1D NMR spectra (1H and/or 13C) of an unknown compound with up to ~19 heavy atoms and need to predict both molecular formula and connectivity without manual structure hypothesis generation.

Use when when building a comprehensive lipid fragment ion library covering all chain composition and positional isomer variants (e.g., 168.6 million entries).

Use when when you have spectroscopic measurements (1D ¹H or ¹³C NMR) that have been decomposed into predicted substructures or fragments, and you need to reconstruct the full molecular formula and connectivity graph.

Use when after feature detection and alignment have produced a feature table with MS/MS spectra, and you have access to a reference spectral database (e.g., xenobiotic reaction libraries or public databases).

Use when you have extracted parallel feature streams from a CNN backbone (local spectral patterns) and a Transformer backbone (global dependencies) in 1H NMR spectra, and you need to fuse them into a single embedding for bi-encoder or cross-encoder processing before compound identification scoring.

Use when when you have 1D NMR spectra (¹H and/or ¹³C) as input and need to predict complete molecular structure (both molecular formula and bond connectivity) for molecules with up to 19 heavy atoms.

Use when you have 1D or 2D NMR spectra (1H and/or 13C) and need to predict unknown molecular structure (formula and connectivity) up to ~19 heavy atoms; or you have a set of molecular fragment-structure pairs and need to model how fragments assemble into complete structures.

Use when you have 2D NMR spectral data from multiple samples (a large sample set) including both heteronuclear experiments (HSQC, HMBC) and homonuclear experiments (COSY), and your goal is to identify which structural features are conserved across samples, prioritize samples by scaffold similarity.

Use when you have identified one or more proton NMR spectral regions-of-interest (ROIs)—defined by lower and upper chemical-shift bounds in ppm—from complex biological samples (serum, saliva, urine, tissue, CSF) and need to generate a ranked list of plausible metabolite identities.

Use when when you need to expose a multi-step spectral processing workflow (parse → extract → render → annotate → compress) as a web service endpoint that accepts jcamp/RAW/mzML spectral files and must return coordinated output artifacts (modified JCAMP files, peak tables, and PNG images) in a.

Use when when your application needs to enrich or predict spectral properties (NMR peaks, molecular structure) by querying external databases, and you have peak data (chemical shift, multiplicity, integration) that must be transformed into a remote service's query format, validated, and the.