AI Agent Workflow Collection

Use when when you have a neural network or machine learning model with multiple tunable hyperparameters (layer size, regularization strength, dropout) or design choices (e.

Use when after running inference on a trained structure prediction model with one or more input modalities (1H NMR, 13C NMR, or combined), you have generated predicted molecular formulas and connectivity graphs that need to be compared against known ground truth structures.

Use when after peak detection and MS1 feature extraction from FIA-MS, GC-MS, LC-MS(/MS), or CE-MS data, when you need to identify unknown metabolites by matching observed m/z values to a reference database and want to recover HMDB identifiers, molecular formulas, and structural annotations for.

Use when you have preprocessed metabolite intensity data (log2-transformed, zero-mean unit-variance standardized) mapped to compound annotations, and you need to derive activity scores for a set of metabolite groups (pathways, Molecular Families, Mass2Motifs, or custom metabolite sets) to rank them.

Use when after computing PLAGE-derived activity scores for pathways or metabolite sets (Molecular Families, Mass2Motifs) from log2-standardized metabolomics intensity data.

Use when after running DESeq() and extracting raw results with results(), when you have log fold change estimates with high variance and wish to improve their precision.

Use when you have an LC-MS peak-intensity matrix with observed m/z values (from negative or positive mode ionization) and need to map each peak to candidate neutral masses in KEGG. Use this skill when you have a curated Cpd.

Use when after computing a histogram of all pairwise mass differences from an MSI dataset, use this skill when you have observed mass difference peaks that may correspond to known adducts (e.g., [M+H]+, [M+Na]+, [M−H2O]+).

Use when you have a trained formula ranking model (such as MIST-CF) and want to measure the specific performance gain from incorporating multiple positive-mode adduct types (e.g., [M+H]+, [M+Na]+, [M+K]+, [M+NH4]+) instead of restricting predictions to [M+H]+ only.

Use when after accurate mass searching has assigned multiple detected m/z features to the same metabolite via positive and negative adduct libraries, and before sample-level feature merging.

Use when you have a characterized lipid species (with defined class and fatty acid composition) and need to predict which adduct forms will ionize under your experimental ionization mode (positive or negative ESI), and you require accurate precursor m/z values for DDA method configuration or.

Use when when initializing an mWISE annotation pipeline with a new or custom KEGG database, or when you need to reconstruct the Cpd.Add matching table with modified adduct/fragment specifications (e.g., subset to instrument-specific adducts or adjust frequency thresholds).

Use when when you have binned mass spectrometry imaging peaks and want to understand which detected mass-to-charge ratios represent the same metabolite in different ionization states (parent vs. adduct form).

Use when when annotating m/z features against a metabolite database (HMDB, Lipidmaps, etc.) and the sample preparation, ionization method, or polarity mode favors specific adduct species.

Use when parsing, standardizing, or filtering MS spectra from mixed or heterogeneous databases where adduct assignment may be manually entered, auto-inferred, or missing.

Use when you have a set of in silico-predicted compounds (with SMILES structures) and an experimental metabolomics peak list (m/z values), and you need to filter predictions to only those that could plausibly be detected.

Use when when you have a metabolite SMILES structure and need to predict which adduct ions will appear in a mass spectrum acquired with a chemical derivatizing matrix.

Use when you have computed a histogram of pairwise mass differences from MS imaging data and need to (1) identify which observed mass differences correspond to biologically relevant or chemically known adducts, or (2) rank the most frequently observed mass differences to discover dominant adduct.

Use when after identifying statistically significant LC-MS features (e.g. via MB-VIP permutation testing) when you need to consolidate redundant measurements of the same metabolite arising from different ionisation adducts (e.g. [M+H]+, [M+Na]+, [M−H]−).

Use when when you have an LC-MS feature table with m/z and retention time columns and need to identify which observed ions correspond to the same neutral compound under different ionization conditions and isotopic enrichment.

Use when ingesting mass spectrometry spectra from heterogeneous databases or libraries where adduct annotations may be incomplete, incorrectly formatted, or inconsistent with the ionization mode. Use it before downstream analysis (e.

Use when you have statistically significant LC-MS features (e.g., filtered by p-value < 0.01) from multi-assay metabolomics datasets and need to group features that represent the same metabolite ionized under different ESI conditions (e.g., [M+H]+, [M+Na]+, [M+NH4]+).

Use when when you have access to annotated MS/MS spectra from a specific ionization mode (e.g., negative ESI) or adduct class (e.

Use when immediately after calling squidpy.gr.spatialneighbors() or similar spatial graph construction methods on an AnnData object. It is essential when validating that the computed spatial graph has been correctly stored in adata.