DuReS: An R Package for Denoising Experimental Tandem Mass Spectra and Metabolite Annotation

Mass spectrometry-based untargeted metabolomics is a
powerful technique
for profiling small molecules in biological samples, yet accurate
metabolite identification remains challenging. The presence of random
noise peaks in tandem mass spectra can lead to false annotations and
necessitate time-consuming manual verification. A common method for
removing noise from mass spectra is intensity thresholding, where
low-intensity peaks are discarded by applying a user-defined cutoff.
However, determining an optimal threshold is often data set-specific
and may still retain many noisy peaks. We hypothesize that true signal
peaks consistently recur across replicate tandem spectra generated
from the same precursor ion, unlike random noise. Here, we present
a freely available R package, Denoising Using Replicate Spectra (DuReS)
(https://github.com/BiosystemEngineeringLab-IITB/dures), which
accepts mzML files and feature lists and returns high-quality annotations
and denoised mzML files, enabling users to integrate the denoising
pipeline into their workflow seamlessly. This package is designed
for data-dependent acquisition mode (DDA) data. It has (i) the main
denoising module and (i) an optional tuning module to determine each
data set’s optimal recurrence frequency cutoff (Fthreshold), considering variations in the intrinsic noise
characteristics. We tested the tool on eight representative data sets
selected from those available in metabolomics repositories. Our approach
minimizes signal loss while maximizing noise reduction, effectively
preserving diagnostically significant low-intensity fragments that
would otherwise be lost through conventional intensity thresholding.
This improves spectral matching metrics, leading to more accurate
annotations and fewer false positives.