Electronic noise ion removal

electronic-noise-ion-removal

Summary

Remove spurious ions arising from instrument artifacts in MS/MS spectra by filtering peaks with identical intensities that exceed a statistical threshold, improving spectral quality for downstream matching and annotation. This addresses electronic noise as a major category of contamination in mass spectrometry data.

When to use

Apply this skill when working with raw MS/MS spectra that contain ions with repeated (identical) intensity values—a hallmark of electronic noise rather than true metabolite fragments. Use it before spectral comparison or denoising search workflows when entropy-based similarity metrics must be reliable, or when high-confidence compound identification is required. The empirical threshold of >4 ions with identical intensities is applicable after validation on reference libraries like NIST23.

When NOT to use

• Input spectra are already electronically cleaned or processed by vendor denoising software.
• Analysis goal does not require high-confidence spectral matching (e.g., exploratory, low-resolution screening).
• Spectrum originates from instruments with negligible electronic noise signatures (verify against instrument/method literature first).

Inputs

• MS/MS peak array (numpy array of shape [n_peaks, 2] with columns [m/z, intensity])
• MSP-formatted spectrum file (containing peaks, precursor_mz, and molecular metadata)
• Single spectrum object or batch of spectra from sd.read_msp()

Outputs

• Denoised peak array with electronic noise ions removed (same format as input)
• Peak array ready for downstream entropy_similarity() or formula_denoising() operations
• Batch of denoised spectra when operating in parallel mode

How to apply

Load a peak array (m/z, intensity pairs) from an MSP file or spectrum object using sd.read_msp(). Pass the peak array to the electronic_denoising() function, which identifies and removes ions having identical intensities that occur more than 4 times in a single spectrum—a threshold determined to be statistically unlikely in real MS/MS data. The function returns a cleaned peak array with electronic noise removed but preserves the chemical signal. Apply this as a first-pass filter before formula_denoising() or spectral_denoising() to ensure clean input for entropy similarity calculations.

Related tools

• spectral_denoising (Python package providing electronic_denoising() function and integrated spectral workflow) — https://github.com/FanzhouKong/spectral_denoising
• ms_entropy (Computes spectral entropy and entropy_similarity() for downstream validation of denoising efficacy)
• numpy (Array manipulation and storage of peak m/z and intensity data)

Examples

import spectral_denoising as sd
import numpy as np
peak = np.array([[48.99, 154.0], [63.01, 265.0], [79.02, 521.0]], dtype=np.float32)
peak_denoised = sd.electronic_denoising(peak)

Evaluation signals

• Output peak array contains fewer peaks than input; confirm that only ions with ≥5 identical intensity counts were removed.
• Entropy similarity between denoised spectrum and clean reference increases compared to raw spectrum (verify using entropy_similarity function).
• No fragment m/z values corresponding to known metabolite losses are removed; validate that chemical signal is preserved by comparing against reference library spectra.
• Precursor m/z and base peak m/z remain unchanged after denoising (electronic noise removal should not alter true fragment ions).
• Spectral entropy value of output spectrum is stable (does not artificially inflate due to peak removal artifacts).

Limitations

• Threshold of >4 identical intensities is empirically derived from NIST23 and may not generalize to all instruments, acquisition methods, or mass ranges.
• Cannot distinguish between electronic noise and legitimate multiply-charged or isobaric fragment ions that coincidentally share intensity values.
• Removal is absolute and irreversible; spectra with naturally clustered intensity distributions (e.g., from isotope patterns) may suffer false-positive filtering.
• Requires peak arrays as input; does not handle raw binary vendor formats (.raw, .d) without prior conversion via standard MS data readers.

Evidence

• [intro] Noise ions in MS/MS spectra are largely categorized as 1. electronic noises and 2. chemical noises.: "Noise ions in MS/MS spectra are largely categorized as 1. electronic noises and 2. chemical noises."
• [other] The electronic_denoising function removes obvious electronic noise ions in MS/MS spectra: "The electronic_denoising function removes obvious electronic noise ions in MS/MS spectra"
• [other] According to empiracally tested on NIST23 database, in a given spectrum, the number of ions with identical intensities more than 4 is extremely unlikely: "in a given spectrum, the number of ions with identical intensities more than 4 is extremely unlikely"
• [other] Apply electronic_denoising() to remove ions with identical intensities >4.: "Apply electronic_denoising() to remove ions with identical intensities >4."
• [readme] peak_denoised = sd.electronic_denoising(peak): "Perform electronic denoising [section=other; evidence='peak_denoised = sd.electronic_denoising(peak)']"