Claude Skills by HolobiomicsLab

Use when when you have a peptide sequence and need to predict which fragment ions (B and Y series) should appear in an MS2 spectrum at a known isotopic abundance (e.g., natural 13C at 1.07%, or stable isotope labeling at 50% enrichment).

Use when you have a peptide sequence, precursor charge state, and observed m/z spectrum and need to generate the expected isotopic distribution to match against observed precursor or fragment ion peaks. This is especially critical in SIP workflows where incorporation probabilities (e.g., 0.

Use when after embedding MS/MS spectra into a 32-dimensional vector space using GLEAMS, when you need to identify and group all spectra originating from the same peptide sequence.

Use when you have a list of polypeptide sequences (one per line or CSV format) and need to compute average mass (weighted by natural isotope abundances) to compare against experimental LC-MS or MS/MS data where the full isotopic distribution—not just the most abundant peak—is relevant for peptide.

Use when when you have an observed MS/MS spectrum and need to annotate fragment peaks against a known modified peptide sequence. Specifically: (1) you possess a peptide amino acid sequence with known or predicted post-translational modifications at specific positions;

Use when you have peptide or protein sequences (as FASTA strings or text identifiers) and need to compute their mass, isotopic envelope, charge state, isoelectric point, or chromatographic retention time for MS matching, peak annotation, or property prediction.

Use when when you have a pre-trained Casanovo model, annotated MS/MS spectra in MGF format, and want to benchmark whether beam search decoding improves peptide prediction quality over the default greedy decoding strategy.

Use when when you have a peptide sequence and need to predict its fragmentation pattern under specific isotopic labeling conditions (e.g., natural 13C abundance at 1%, or enriched 13C at arbitrary levels).

Use when when you have a peptide sequence, observed MS2 spectrum peaks (m/z, intensity, charge state), and need to determine which theoretical fragment ions (B and Y ions) match the observed data—particularly in stable isotope probing (SIP) experiments where peptides carry heavy isotope labels (e.

Use when when you have a tandem mass spectrum (MSMS) with known peptide sequence and wish to assess whether enabling neutral loss annotation (e.g., NH3: −17.026549, H2O: −18.010565) increases the proportion of observed m/z peaks that can be matched to predicted fragment ions.

Use when when you have high-resolution tandem mass spectrometry data (in mzML or mzXML format) and a spectral library, and need to identify peptides including those with post-translational modifications.

Use when you have a collection of MS/MS spectra (in mzML or MGF format) and need to group or compare spectra from the same peptide without prior sequence annotation.

Use when when you have a tandem mass spectrometry spectrum with a known or inferred peptide sequence that may contain post-translational modifications (phosphorylation, glycosylation, cross-links), and you need to annotate which observed m/z peaks correspond to specific fragment ion types (b, y, a.

Use when after a transformer-based de novo sequencing model (such as Casanovo) generates candidate peptide sequences from MS/MS spectra, before exporting results or using them in database matching or visualization workflows.

Use when when you have aligned and quantified mass spectrometry features from multiple natural extracts (via MZmine2/3), paired with in silico annotation results (ISDB or SIRIUS), and you need to prioritize samples for chemical discovery based on the proportion of unannotated, extract-specific.

Use when after LC-MS data acquisition is complete (or during real-time monitoring) and you have loaded processed LC-MS data in mzML or vendor format and defined QC criteria (retention time windows, m/z tolerances, intensity thresholds) for your internal standards and target analytes.

Use when you have raw MS2 spectra from a sample and need to collapse them into a single sample-level representation for comparison across multiple samples, particularly when samples have poor feature overlap, strong retention time shifts between LC methods, or were acquired on different mass.

Use when you have a raw or pre-processed LC-MS feature table with multiple samples and need to normalize for inter-sample intensity biases before downstream statistical analysis.

Use when you have computed multiple independent scoring functions (e.g., standardised strain correlation and IOKR) for a large set of potential genomic–metabolomic links and wish to identify subsets enriched for validated links.

Use when you have implemented or reconstructed a performance-critical computational module (e.g., an expeditious querying engine, a database lookup accelerator, or a real-time matching algorithm) and need to validate that it achieves claimed throughput targets.

Use when when a trained model produces probabilistic or ensemble predictions and you need to achieve a specific target accuracy metric (e.g., RMSE ≤ 0.1) or minimize error on a test set, but the unfiltered model does not meet that target.

Use when after computing Multi-Block Variable Importance in Projection (MB-VIP) scores on a fitted MB-PLS discriminant model, when you need to distinguish signal features from noise by establishing empirical significance thresholds rather than relying on parametric assumptions.

Use when you have a feature list (containing m/z, retention time, and molecular formula or neutral mass per feature) extracted from LC- or GC-HRMS data in mzML format with data-dependent acquisition, and you need to rank or flag features as likely PFAS compounds to reduce manual review burden in.

Use when you have centroided data-dependent acquisition (ddMS2) mzML spectra from LC- or GC-HRMS and need to prioritize putative PFAS features.

Use when you have centroided MS2 spectra (ddMS2 data in mzML format) from HRMS analysis and need to identify potential PFAS compounds among thousands of features.

Use when when you have a feature list (m/z, retention time, intensity) from LC- or GC-HRMS non-target screening and need to identify PFAS-like homologous series characterized by CF₂ (14 Da) repeating units.

Use when you have detected features in LC- or GC-HRMS data (via pyOpenMS or custom feature tables) and need to systematically rank them for likelihood of being PFAS compounds.

Use when you have an m/z-resolved feature list from LC- or GC-HRMS analysis (either detected by pyOpenMS or provided as a custom Excel table) and need to prioritize potential PFAS compounds by identifying clusters of homologous structures.

Use when after generating a Chemical Feature Tree from q2-qemistree (or any tree artifact) and before using it for alpha-diversity or beta-diversity phylogenetic analyses.

Use when you have one or more peptide or protein sequences in string format and need to calculate their mass, isotopic envelope (m/z and intensity pairs), charge state behavior, isoelectric point, or predicted chromatographic retention.

Use when you have a published computational pipeline with deposited code and validation data, and you need to verify that the pipeline can be executed end-to-end to reproduce reported validation metrics (annotation accuracy, coverage, or equivalent performance benchmarks).

Use when before launching the DaDIA metabolomics pipeline or any multi-package workflow, when you have an R environment with potentially mixed or unknown package versions and need to confirm that R ≥4.0, XCMS ≥3.11.4, metaMS ≥1.25.

Use when after generating a numerical visualization (e.g., confusion matrix, heatmap, or similarity array) using matplotlib, when you need to add axis labels, class names, colormaps, normalization annotations, colorbars, titles, and export the figure in a publication-ready format (PNG or PDF).

Use when when running NPLinker in PODP mode (as opposed to local mode), you need to fetch and validate project metadata from PODP, orchestrate downloads of GNPS molecular networking data, AntiSMASH BGC predictions, BigScape clustering results, and MIBiG reference metadata, then organize them into.

Use when when performing targeted peak detection on LC-MS data where compounds have been assigned expected ionization polarities (positive or negative mode) in the target list, and you want to prevent false peak assignments from the opposite polarity and avoid manual pre-filtering of raw data by.

Use when you have a comprehensive target list (containing compounds from both positive and negative ionization modes) but need to screen or detect peaks in a single LC-MS run acquired in a specific polarity mode.

Use when after fitting a polynomial calibration model to tunemix reference data in DEIMoS, assess whether the model explains sufficient variance in the m/z–drift-time–CCS relationship.

Use when preparing mass spectrum input tensors for transformer encoder layers in IDSL_MINT.

Use when you have LC-MS metabolomics data in positive ionization mode and have already performed XCMS feature detection and RAMClustR clustering.

Use when when training a formula rescoring model on MS/MS spectra where positive examples (correct molecular formulas) are unevenly distributed across molecular formula groups or vastly outnumbered by negative examples (incorrect candidates), resulting in class imbalance that degrades model.

Use when when you have tandem mass spectrometry data (LC-MS/MS in MGF, mzML, mzXML, or mzData format) paired with either genome sequences or precursor peptide predictions, and you want to confirm the presence and identity of modified ribosomally synthesized and post-translationally modified.

Use when you have centroided LC-MS/MS spectra (in MGF, mzXML, mzML, or mzData format) and genomically-predicted precursor peptide sequences, and you need to identify which predicted RiPPs are actually expressed and modified in the sample.

Use when you have raw GC–MS or LC–MS data in two-dimensional m/z vs retention time format and need to identify marker features at parts-per-billion sensitivity without relying on conventional peak picking.

Use when when annotating m/z features from Cardinal MSImagingExperiment objects or LC-MS datasets against metabolite databases (HMDB, Lipidmaps) and you need to exclude matches where the mass difference exceeds your instrumental accuracy or analysis tolerance.

Use when you have extracted a list of candidate molecular formulae for a given m/z value and need to rank them by plausibility.

Use when you have a small training dataset for molecular property prediction (e.g., <500 samples from PredRet or MoNA databases) and a pre-trained GNN model is available that was trained on a related, larger molecular corpus.

Use when you have computed spectral similarity scores from multiple methods (e.g., MS2DeepScore, Spec2Vec, modified cosine) on the same set of spectrum pairs, have assigned ground-truth structural similarity labels (e.

Use when when you have computed similarity scores (e.g., MS2DeepScore, Spec2Vec, modified Cosine) between pairs of spectra or compounds and want to compare their ability to retrieve chemically related pairs. Apply this skill if you have ground-truth structural similarity labels (e.

Use when you have extracted fragmentation patterns from a collection of MS/MS spectra (using mineMS2) and have partitioned spectra into components via GNPS molecular networking (e.g., connected components, cliques, or high-similarity pairs with cosine > threshold).

Use when you have loaded raw tandem MS spectra (in MGF, mzML, or similar format) and need to prepare them for peptide identification, spectral library matching, or intensity-based analysis.