agentsclimarketplace

Reactant ion peak exclusion

Skill HolobiomicsLab/asb-skill-collections/collections/metabolomics/v2/skills/reactant-ion-peak-exclusion

Use when when performing peak detection on Gas Chromatography–Ion Mobility Spectrometry samples where the Reactant Ion Peak (a high-intensity background signal from the ion source) would otherwise be misidentified as an analyte peak, causing false positives in the peak list.From its SKILL.md

Install
npx -y skills add HolobiomicsLab/asb-skill-collections --skill reactant-ion-peak-exclusion

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 15 stars15 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its file declares

Copied from the file, not written here

The file declares its own license as CC-BY-4.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

5.7 KB, 892 tokens by cl100k_base, as published. Nobody here has run it

reactant-ion-peak-exclusion

License: restricted — no clear open-source license detected for the underlying tool; verify licensing before commercial use or redistribution. <!-- asb-license-banner -->

Summary

Suppress detection of Reactant Ion Peaks (RIP) during GCIMS peak detection to avoid false positives from instrument noise. RIP exclusion is a filtering parameter applied during the CWT-based findPeaks step to ensure only true analyte peaks are retained in the peak_list output.

When to use

When performing peak detection on Gas Chromatography–Ion Mobility Spectrometry samples where the Reactant Ion Peak (a high-intensity background signal from the ion source) would otherwise be misidentified as an analyte peak, causing false positives in the peak list.

When NOT to use

  • When RIP signal itself is the analytical target (e.g., studying ion source behavior or instrument diagnostics)
  • When working with datasets where RIP has already been removed by hardware filtering or preprocessing steps
  • On non-ion-mobility or non-GCIMS data where the concept of a Reactant Ion Peak does not apply

Inputs

  • aligned GCIMSDataset object (post-alignment in both drift time and retention time)
  • CWT algorithm parameters (noise_level, exclude0scaleAmpThresh, extension factors)

Outputs

  • peak_list table with RIP-free detected peaks (columns: retention time, drift time, intensity)

How to apply

Set the exclude_rip parameter to TRUE when invoking the findPeaks function with CWT algorithm on an aligned GCIMSDataset object. This suppresses RIP detection during the peak-finding phase across all samples in the dataset. The RIP typically appears as a strong signal at characteristic drift and retention times and, without exclusion, would dominate the peak list and interfere with downstream peak clustering and compound identification. Apply this setting consistently across all samples to ensure comparable peak detection.

Related tools

  • GCIMS (R package providing the findPeaks function with exclude_rip parameter) — https://github.com/sipss/GCIMS
  • R (Execution environment for GCIMS and peak detection workflow)

Examples

library(GCIMS); peak_list <- findPeaks(aligned_dataset, exclude_rip = TRUE, noise_level = <optimized_value>)

Evaluation signals

  • Peak list does not contain high-intensity signals at the characteristic RIP retention and drift time coordinates
  • Number of detected peaks is reduced compared to peak detection without RIP exclusion (exclude_rip = FALSE)
  • Downstream peak clustering results in distinct, chemically interpretable compound clusters without a dominant high-intensity outlier
  • Per-sample peak intensity distributions show removal of the characteristic RIP intensity spike
  • Output peak_list schema validation passes with all required columns (retention time, drift time, intensity) and no unexpected RIP markers

Limitations

  • RIP exclusion requires prior knowledge or empirical determination of RIP position in drift and retention time space; misidentification of RIP coordinates can lead to loss of genuine analyte peaks
  • Effectiveness depends on proper prior alignment of the dataset in both drift time and retention time; misaligned data may place RIP at varying positions, reducing exclusion effectiveness
  • Cannot distinguish RIP from very early-eluting, high-abundance analytes if they co-localize in drift–retention time space; user must validate peak assignments

Evidence

  • [intro] Apply the findPeaks function with CWT algorithm and documented noise_level parameter optimization on the first sample to determine optimal settings.: "Apply the findPeaks function with CWT algorithm and documented noise_level parameter optimization"
  • [intro] Execute findPeaks across all samples in the aligned dataset using the optimized CWT parameters and exclude_rip = TRUE to suppress Reactant Ion Peak detection.: "Execute findPeaks across all samples in the aligned dataset using the optimized CWT parameters and exclude_rip = TRUE to suppress Reactant Ion Peak detection"
  • [intro] Exclude RIP (Reactant Ion Peak) during peak detection: "Exclude RIP (Reactant Ion Peak) during peak detection"
  • [intro] Generate the per-sample peak_list table as structured output containing detected peaks with retention time, drift time, and intensity annotations.: "Generate the per-sample peak_list table as structured output containing detected peaks with retention time, drift time, and intensity annotations"

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.