Molecular Formula Isotopic Profile Matching

Molecular Formula Annotation via Isotopic Profile Matching

Summary

Assign molecular formulas to detected LC/HRMS peaks using only MS1 data by comparing observed isotopic patterns against theoretically predicted patterns for candidate formulas from a prioritized chemical space. This approach leverages natural isotope abundance to rank and select the best-matching formula for each peak without requiring MS2 spectra.

When to use

Use this skill when you have detected peaks from untargeted LC/HRMS analysis (via IDSL.IPA or equivalent peak picker) with m/z and retention time values, and you need to assign molecular formulas to those peaks. Apply it when MS1 isotopic profile data are available and high-resolution mass accuracy (typically <5 ppm) is achievable, but MS2 spectra are unavailable, limited, or you want to complement MS2-based annotation. Particularly valuable for exposomics and large population-scale studies (n > 500) where diverse molecular formula sources beyond common metabolite databases are expected.

When NOT to use

• Input data already have high-confidence structural identifications from MS2 spectra or other orthogonal methods—molecular formula matching may be redundant.
• MS1 mass resolution is insufficient (typically <5 ppm, and degraded by instrument limitations or data processing)—isotopic pattern matching will have poor discriminatory power.
• Peak detection has already been aggregated or merged without preserving individual m/z and intensity lists—isotopic profile reconstruction requires the full scan-level data.

Inputs

• Peaklist from IDSL.IPA or equivalent peak picker (m/z, intensity, retention time values)
• MS1 raw mass spectrometry data (mzXML, mzML, or netCDF format)
• Pre-calculated isotopic profile database (IPDB) in Rdata format for target chemical space and ionization mode
• UFA parameter spreadsheet specifying HRMS data location, peak alignment directory, IPDB path, and output directory

Outputs

• Annotated peaklist table with original m/z, intensity, assigned molecular formula, isotopic matching score, and formula rank
• Aggregated molecular formula assignments on aligned peak table (unique to IDSL.UFA)
• Batch untargeted isotopic profile match figures visualizing observed vs. predicted patterns

How to apply

Load the peaklist (m/z, intensity, retention time values) generated by IDSL.IPA into R and initialize IDSL.UFA with a pre-calculated isotopic profile database (IPDB) corresponding to your chemical space of interest and ionization mode (positive or negative). For each detected peak, extract the observed MS1 isotopic profile (the monoisotopic m/z and its associated isotope peaks at M+1, M+2, etc.). Compare this observed pattern against theoretical isotopic profiles for candidate molecular formulas using combinatorial IUPAC atomic mass rules. Rank candidate formulas by isotopic profile similarity (based on the match between predicted and observed relative isotope abundances and m/z positions). Assign the highest-ranking formula to each peak and record the isotopic matching score and rank. Aggregate results across the aligned peak table to capture recurring formula assignments. Parallel processing and user-friendly parameter spreadsheets (UFA_parameters.xlsx) guide mass accuracy tolerance, isotope matching thresholds, and aggregation rules.

Related tools

• IDSL.UFA (Primary R package that performs isotopic profile matching and molecular formula annotation on IDSL.IPA peaklists) — https://github.com/idslme/IDSL.UFA
• IDSL.IPA (Upstream peak detection and chromatographic processing tool that generates input peaklists and m/z-RT annotations required by IDSL.UFA) — https://github.com/idslme/IDSL.IPA
• R (Computing environment in which IDSL.UFA runs)
• IDSL.CSA (Optional downstream workflow for clustering recurring ions and generating composite spectra from IDSL.UFA-annotated results) — https://github.com/idslme/IDSL.CSA

Examples

library(IDSL.UFA)
UFA_workflow("path/to/UFA_parameters.xlsx")

Evaluation signals

• Isotopic matching score is high (well-described in UFA_parameters.xlsx; typically normalized 0–1 or similar) and formula ranks are consistent across biological replicates or technical replicates.
• Assigned molecular formulas satisfy chemical feasibility rules (e.g., Hill system element order, valence, DBE constraints) and match known compound databases or expected metabolite/exogenous chemical lists for the study cohort.
• Observed m/z values for the monoisotopic and isotope peaks align with predicted patterns within specified mass accuracy tolerance (typically <5 ppm for high-resolution instruments).
• Aggregated formula assignments on the aligned peak table show low formula redundancy and high uniqueness, indicating that peaks with distinct m/z and retention time map to distinct formulas.
• Visual inspection of batch isotopic profile match figures confirms predicted and observed isotopic envelopes overlap and rank-1 assignments appear biochemically or chemically reasonable for the study context.

Limitations

• Isotopic profile matching relies on accurate, high-resolution MS1 data; poor mass accuracy, low signal-to-noise, or severe peak overlap will degrade formula assignment quality.
• The chemical space of candidate formulas is defined by the IPDB; formulas outside that space (e.g., rare elements, unusual elemental compositions) cannot be detected.
• IDSL.UFA requires prior peak detection and m/z-RT annotation by IDSL.IPA or equivalent; errors in peak picking propagate to formula assignment.
• No changelog is publicly available for IDSL.UFA, limiting reproducibility tracking and version-dependent behavior assessment.
• The approach works best for carbon-containing small organic molecules; inorganic species or very large polymers may have isotopic patterns poorly predicted by IUPAC combinatorial rules.

Evidence

• [other] IDSL.UFA is an R package that annotates peaklists from IDSL.IPA with molecular formulas from a prioritized chemical space using an isotopic profile matching approach, requiring only MS1 data.: "annotate peaklists from the IDSL.IPA package with molecular formula of a prioritized chemical space using an isotopic profile matching approach. The IDSL.UFA pipeline only requires MS1"
• [other] For each detected peak, perform isotopic profile matching by comparing the observed MS1 isotopic pattern against predicted patterns for candidate molecular formulas in the chemical space.: "For each detected peak, perform isotopic profile matching by comparing the observed MS1 isotopic pattern against predicted patterns for candidate molecular formulas in the chemical space."
• [other] Rank candidate formulas by isotopic profile similarity and assign the top-ranked formula to each peak.: "Rank candidate formulas by isotopic profile similarity and assign the top-ranked formula to each peak."
• [readme] The isotopic pattern for a chemical structure can be accurately predicted using a set of combinatorial rules that uses atomic mass tables provided by the International Union of Pure and Applied Chemistry (IUPAC).: "The isotopic pattern for a chemical structure can be accurately predicted using a set of combinatorial rules that uses atomic mass tables provided by the International Union of Pure and Applied"
• [readme] Generating comprehensive in-silico theoretical libraries (known as IPDB) using natural isotopic distribution profiles.: "Generating comprehensive in-silico theoretical libraries (known as IPDB) using natural isotopic distribution profiles"
• [readme] Aggregating annotated molecular formulas on the aligned peak table. This is a very unique feature that is only presented by IDSL.UFA.: "Aggregating annotated molecular formulas on the aligned peak table. This is a very unique feature that is only presented by IDSL.UFA."
• [readme] Analyzing population size untargeted studies (n > 500).: "Analyzing population size untargeted studies (n > 500)"