Agent skill
metabolomics
Metabolomics-specific analysis strategies and domain knowledge
Install this agent skill to your Project
npx add-skill https://github.com/justaddcoffee/open-science-skills/tree/main/metabolomics
SKILL.md
Metabolomics Analysis
When to Use This Skill
- When data contains metabolite measurements
- When analyzing metabolic pathways or fluxes
- When interpreting biochemical mechanisms
Core Concepts
Metabolite Naming
Metabolites have multiple naming conventions:
- IUPAC names: Chemical nomenclature (e.g., "2-aminoethanesulfonic acid")
- Common names: Biology names (e.g., "Taurine")
- Abbreviations: Shorthand (e.g., "Tau")
Always verify metabolite identity before interpreting results.
Pathway Context
Metabolites exist in biochemical pathways:
- Substrates → Enzymes → Products
- Changes in one metabolite affect connected metabolites
- Pathway analysis is more informative than individual metabolites
Example pathway:
Glucose → (HK) → G6P → (G6PDH) → 6PG
If G6P ↑ and 6PG unchanged → suggests bottleneck at G6PDH enzyme
Flux vs Concentration
Concentration: Amount of metabolite present Flux: Rate of metabolite conversion
Key insight:
- High concentration + low downstream product = bottleneck (slow flux)
- Low concentration + high downstream product = high flux
- Calculate flux proxies using ratios: Product/Substrate
Common Metabolomics Patterns
Pattern 1: Substrate Depletion
Precursor ↓↓, Product ↑↑
→ Interpretation: Active consumption, increased flux
Pattern 2: Bottleneck
Substrate ↑↑, Product ↓↓ or unchanged
→ Interpretation: Enzymatic bottleneck, blocked conversion
Pattern 3: Pathway Shutdown
All pathway metabolites ↓↓
→ Interpretation: Reduced pathway activity
Pattern 4: Salvage vs De Novo
De novo intermediates ↓, Salvage products ↑
→ Interpretation: Metabolic shift to energy-efficient salvage
Analysis Strategies
1. Pathway Enrichment
When: You have many differentially abundant metabolites
How:
# Group metabolites by pathway
pathway_metabolites = {
"Glycolysis": ["Glucose", "G6P", "F6P", "FBP", ...],
"TCA Cycle": ["Citrate", "Isocitrate", "α-KG", ...],
"Purine Metabolism": ["AMP", "ADP", "ATP", "IMP", ...]
}
# Count hits per pathway
for pathway, metabolites in pathway_metabolites.items():
hits = [m for m in significant_metabolites if m in metabolites]
enrichment_score = len(hits) / len(metabolites)
Resources:
- KEGG pathways: https://www.genome.jp/kegg/pathway.html
- BioCyc: https://biocyc.org/
2. Flux Index Calculation
When: You want to infer enzymatic activity
How:
# Simple flux proxy: Product / Substrate
flux_index = data["Product"] / data["Substrate"]
# Compare across groups
t_test(flux_index[group1], flux_index[group2])
Common indices:
- Glycolysis flux: FBP / G6P
- TCA flux: Citrate / Acetyl-CoA (if available)
- Salvage flux: Product / Precursor
3. Energy Charge Calculation
When: Assessing cellular energy state
Formula:
# Adenylate energy charge
AEC = (ATP + 0.5*ADP) / (ATP + ADP + AMP)
# Range: 0 (depleted) to 1 (high energy)
# Similar for GTP, CTP, UTP
Interpretation:
- AEC > 0.8: High energy state
- AEC < 0.5: Energy crisis
4. Redox State Assessment
When: Investigating oxidative stress or metabolic state
Ratios:
NAD_ratio = NAD+ / NADH # High = oxidized state
NADP_ratio = NADP+ / NADPH # High = oxidative stress
GSH_ratio = GSH / GSSG # Low = oxidative stress
Metabolomics-Specific Hypotheses
Template Hypotheses
H1: Pathway Shift Hypothesis
"Condition X shifts metabolism from [pathway A] to [pathway B]
due to [mechanism], evidenced by [metabolite pattern]"
H2: Enzymatic Bottleneck Hypothesis
"Enzyme [E] activity is reduced in condition X, causing accumulation
of substrate [S] and depletion of product [P]"
H3: Cofactor Limitation Hypothesis
"Limited availability of cofactor [C] constrains pathway [P],
causing metabolite pattern [M]"
H4: Energy State Hypothesis
"Condition X induces low-energy state, triggering metabolic
reprogramming to salvage pathways"
Literature Search Strategies
Effective Search Queries
For pathway context:
"[metabolite] metabolism pathway"
"[metabolite] biosynthesis regulation"
For mechanistic insights:
"[condition] [metabolite] mechanism"
"[enzyme] regulation [condition]"
For flux studies:
"[pathway] flux analysis"
"[metabolite] turnover rate"
Key Databases
- KEGG: Pathway maps and enzyme info
- HMDB: Human Metabolome Database
- PubChem: Chemical structures and properties
- MetaboAnalyst: Analysis tools and pathway info
Common Pitfalls
❌ Assuming directionality
- Many reactions are reversible
- Check enzyme and equilibrium constants
❌ Ignoring compartmentalization
- Metabolites exist in different cellular compartments
- Mitochondrial vs cytoplasmic pools may differ
❌ Overinterpreting single metabolites
- Always consider pathway context
- One metabolite change can have multiple explanations
❌ Confusing correlation with regulation
- Co-regulation doesn't mean direct interaction
- Use pathway knowledge to infer relationships
❌ Forgetting isomers
- Many metabolites have isomers (e.g., leucine/isoleucine)
- Mass spec may not distinguish them
Quality Checks
Before interpreting results, verify:
- Metabolite identifications are confident (not just m/z matches)
- Normalization was appropriate (sample weight, protein, etc.)
- Missing values handled correctly
- Batch effects addressed
- Biological replicates have reasonable variance
Example Analysis Flow
Observation: ATP levels decreased 30% (p=0.01)
Step 1: Check related metabolites
# Check adenylate pool
print(data[["ATP", "ADP", "AMP"]])
Step 2: Calculate energy charge
AEC = (ATP + 0.5*ADP) / (ATP + ADP + AMP)
Step 3: Search literature
search_pubmed("[condition] ATP depletion mechanism")
Step 4: Generate hypotheses
- H1: Increased energy demand (check ATP consumers)
- H2: Reduced ATP synthesis (check TCA metabolites)
- H3: ATP degradation (check breakdown products)
Step 5: Test hypotheses
# H2: Check TCA cycle metabolites
tca_metabolites = ["Citrate", "Isocitrate", "α-KG", "Succinate", "Fumarate", "Malate"]
test_pathway(tca_metabolites, group_var)
Key Principle
Metabolism is a network, not a list.
Single metabolite changes are clues, not answers. Build mechanistic models by connecting metabolites through known biochemical pathways.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
structural-biology
Structural biology analysis including protein structure validation, AlphaFold interpretation, and structural comparisons
genomics
Genomics and transcriptomics analysis strategies
data-science
Statistical analysis strategies and data exploration techniques
obsidian-vault
Search, create, and manage notes in the Obsidian vault with wikilinks and index notes. Use when user wants to find, create, or organize notes in Obsidian.
edit-article
Edit and improve articles by restructuring sections, improving clarity, and tightening prose. Use when user wants to edit, revise, or improve an article draft.
handoff
Compact the current conversation into a handoff document for another agent to pick up.
Didn't find tool you were looking for?