Targeted metabolomics quantifies a closed, pre-defined panel of known metabolites and reports absolute concentrations with units, using MRM/SRM on a triple quadrupole or PRM on a high-resolution instrument. The central enemy is the matrix effect: co-eluting matrix suppresses analyte ionization, and only a co-eluting stable-isotope-labeled internal standard truly corrects it. This guide covers building a calibration assay, choosing the internal-standard and weighting strategy, confirming identity by ion ratio, and validating to a depth that matches the decision the number supports.
# Skyline (free, vendor-neutral) for transition lists, integration, export
# Download from: https://skyline.ms/
# R for post-export curve fitting and validation metrics
Rscript -e 'install.packages(c("ggplot2", "dplyr"))'
# Optional Python path
pip install pandas numpy scipy
Conceptual prerequisites: an authentic reference standard for each analyte (its purity scales every reported number), ideally one stable-isotope-labeled internal standard per analyte, a defined transition list (quantifier + qualifier per analyte with collision energies), and a decision about how much validation the application demands.
Tell your AI agent what you want to do: - "Build a 1/x^2-weighted calibration curve and accept it by back-calculated %RE, not R-squared" - "Normalize each analyte to its co-eluting internal standard and quantify the samples" - "Confirm identity with the quantifier/qualifier ion ratio and flag interferences" - "Compute LOD and LLOQ and mark samples below the validated range as not reportable" - "Help me decide an internal-standard and validation strategy for a clinical assay"
"Fit unweighted, 1/x, and 1/x^2 calibration curves and pick the weighting that minimizes low-end back-calculated relative error." "Show the per-level %RE for each calibrator and tell me whether the curve passes ICH M10." "Set the LLOQ to the lowest calibrator within +/-20% accuracy."
"Normalize analyte areas to the SIL-IS and back-calculate concentrations from the response-ratio curve." "Apply the dilution factor and report concentrations in micromolar, flagging anything below the LLOQ."
"Compute the qualifier/quantifier ion ratio per sample and flag any outside +/-30% of the calibrator ratio." "Check carryover in the blank injected after the top calibrator." "Estimate the matrix factor and the IS-normalized matrix factor across matrix lots."
"My panel has 40 chemically diverse metabolites and one global IS -- where is my accuracy at risk?" "Lay out the ICH M10 parameters I need for a regulated PK assay versus an exploratory study." "Should I use a deuterated or 13C internal standard, and what do I verify before trusting it?"