agentleFS
dailycafidailycafi/metabolism-skills/skills/metabolomics-analysis/metabolite-annotation/SKILL.md

bio-metabolomics-metabolite-annotation

Skill7 starsChanged 8 months agoMITView on GitHub

Annotate and identify metabolomics features by matching m/z, retention time, and MS/MS spectra against databases. Use when: user has a feature table and wants compound IDs, needs to annotate m/z values, assign metabolite identities with confidence levels, or match features against HMDB/METLIN/MassBank. Triggers: identify features, m/z annotation, compound identification, metabolite ID, putative annotation, MSI confidence levels, adduct matching, ppm tolerance, neutral mass search, what compound is this m/z.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before using code patterns, verify installed versions match.

    line 16
  2. Distinctive:

    If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API …

    line 21

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Version Compatibility

Reference examples tested with: pandas 2.2+, xcms 4.0+

Before using code patterns, verify installed versions match. If versions differ:

  • Python: pip show <package> then help(module.function) to check signatures
  • R: packageVersion('<pkg>') then ?function_name to verify parameters
  • CLI: <tool> --version then <tool> --help to confirm flags

If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.

Metabolite Annotation

Database Matching by m/z

Goal: Generate putative metabolite identifications by matching observed m/z values against HMDB.

Approach: Convert m/z to neutral mass by subtracting adduct mass, then query HMDB within a specified ppm tolerance.

"Annotate my metabolomics features with compound identities" → Match detected features against metabolite databases by exact mass, MS/MS spectra, and retention time to assign compound identities with confidence levels.

library(MetaboAnalystR)

# Load feature table
features <- read.csv('feature_table.csv')

# Search HMDB by exact mass
search_hmdb <- function(mz, adduct = '[M+H]+', ppm = 10) {
    # Calculate neutral mass from m/z
    adduct_masses <- list(
        '[M+H]+' = 1.007276,
        '[M+Na]+' = 22.989218,
        '[M-H]-' = -1.007276,
        '[M+Cl]-' = 34.969402
    )

    neutral_mass <- mz - adduct_masses[[adduct]]

    # Query HMDB (or local database)
    # Returns putative matches
    matches <- QueryHMDB(neutral_mass, ppm)
    return(matches)
}

# Apply to all features
annotations <- lapply(features$mz, function(m) search_hmdb(m, '[M+H]+', 10))

MS/MS Spectral Matching

from matchms import calculate_scores
from matchms.importing import load_from_mgf
from matchms.similarity import CosineGreedy

# Load query spectra
queries = list(load_from_mgf('sample_msms.mgf'))

# Load reference library (e.g., GNPS, MassBank)
references = list(load_from_mgf('reference_library.mgf'))

# Calculate similarity scores
similarity = CosineGreedy(tolerance=0.01)
scores = calculate_scores(references, queries, similarity)

# Get best matches
for query_idx, query in enumerate(queries):
    best_match_idx = scores.scores[:, query_idx].argmax()
    best_score = scores.scores[best_match_idx, query_idx]

    if best_score > 0.7:
        ref = references[best_match_idx]
        print(f'{query.get("precursor_mz")}: {ref.get("compound_name")} (score={best_score:.2f})')

SIRIUS + CSI:FingerID

# Molecular formula and structure prediction
sirius \
    --input sample.ms \
    --output sirius_results \
    --database hmdb \
    formula \
    fingerid

# Output structure:
# sirius_results/
#   compound_1/
#     formula_candidates.tsv
#     fingerid_candidates.tsv

MetFrag In Silico Fragmentation

library(metfRag)

# Configure MetFrag search
settings <- list(
    DatabaseSearchRelativeMassDeviation = 10,
    FragmentPeakMatchAbsoluteMassDeviation = 0.01,
    FragmentPeakMatchRelativeMassDeviation = 10,
    MetFragDatabaseType = 'HMDB',
    NeutralPrecursorMass = 147.0532
)

# Run fragmentation prediction
results <- run.metfrag(settings, spectrum_file = 'query_spectrum.txt')

RT Prediction for Validation

from deepchem.models import GraphConvModel
import pandas as pd

# Use predicted RT to validate annotations
# Compare observed RT with predicted RT from chemical structure

def validate_annotation(observed_rt, smiles, rt_model):
    '''Check if observed RT matches prediction'''
    predicted_rt = rt_model.predict(smiles)
    rt_error = abs(observed_rt - predicted_rt)

    if rt_error < 30:  # seconds
        return 'confident'
    elif rt_error < 60:
        return 'probable'
    else:
        return 'unlikely'

Confidence Levels (MSI)

# Metabolomics Standards Initiative levels
assign_confidence <- function(annotation) {
    if (!is.null(annotation$authentic_standard)) {
        return(1)  # Identified by authentic standard
    } else if (!is.null(annotation$msms_match) && annotation$msms_score > 0.8) {
        return(2)  # MS/MS match to database
    } else if (!is.null(annotation$formula_match)) {
        return(3)  # Formula confirmed
    } else if (!is.null(annotation$mass_match)) {
        return(4)  # Mass match only
    } else {
        return(5)  # Unknown
    }
}

# Apply to annotations
features$confidence_level <- sapply(annotations, assign_confidence)

CAMERA Adduct Annotation

library(CAMERA)

# Identify adduct and isotope patterns
xsa <- xsAnnotate(xcms_set)
xsa <- groupFWHM(xsa, perfwhm = 0.6)
xsa <- findIsotopes(xsa, mzabs = 0.01, ppm = 10)
xsa <- findAdducts(xsa, polarity = 'positive',
                   rules = c('[M+H]+', '[M+Na]+', '[M+K]+', '[M+NH4]+'))

# Get annotated features
annotated <- getPeaklist(xsa)
annotated$adduct  # Adduct assignment
annotated$isotopes  # Isotope group
annotated$pcgroup  # Correlation group

Batch Annotation Pipeline

library(tidyverse)

annotate_features <- function(feature_table, ppm = 10, polarity = 'positive') {
    results <- feature_table %>%
        rowwise() %>%
        mutate(
            # Calculate possible neutral masses
            mass_h = ifelse(polarity == 'positive', mz - 1.007276, mz + 1.007276),

            # Query databases
            hmdb_match = list(query_hmdb(mass_h, ppm)),
            kegg_match = list(query_kegg(mass_h, ppm)),

            # Best match
            best_match = get_best_match(hmdb_match, kegg_match),
            compound_name = best_match$name,
            compound_id = best_match$id,
            mass_error_ppm = (abs(mz - best_match$mz) / mz) * 1e6
        )

    return(results)
}

# Example query functions (implement based on your database access)
query_hmdb <- function(mass, ppm) {
    # Query HMDB API or local database
    # Return list of matches with name, id, formula, mass
}

Export Annotated Results

# Create annotation report
annotation_report <- features %>%
    select(feature_id, mz, rt, compound_name, compound_id,
           formula, confidence_level, mass_error_ppm, adduct) %>%
    arrange(confidence_level, desc(intensity))

write.csv(annotation_report, 'annotated_features.csv', row.names = FALSE)

# Summary
cat('Annotation summary:\n')
cat('  Level 1 (confirmed):', sum(annotation_report$confidence_level == 1), '\n')
cat('  Level 2 (MS/MS match):', sum(annotation_report$confidence_level == 2), '\n')
cat('  Level 3 (formula):', sum(annotation_report$confidence_level == 3), '\n')
cat('  Level 4 (mass only):', sum(annotation_report$confidence_level == 4), '\n')
cat('  Unknown:', sum(annotation_report$confidence_level == 5), '\n')

Related Skills

  • xcms-preprocessing - Generate feature table
  • pathway-mapping - Map annotated metabolites to pathways
  • proteomics/spectral-libraries - Similar spectral matching concepts

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.