Use of Ion Oscillation Phase Values for Peak Classification and Noise Filtering Efficiency Estimation in Orbitrap...

An unprocessed high-resolution mass spectrum of natural organic matter contains tens of thousands of peaks, many of which are uninformative and associated with noise. Several noise filtering algorithms have been proposed in an attempt to remove them, but there is still no universal solution. Therefore, given the rapid development of high-resolution mass spectrometry and its increasing use in various scientific fields, it is essential to improve existing noise filtering algorithms and develop new ones. Previously proposed solutions have used peak properties such as intensity, mass-to-charge ratio, or peak reproducibility. In this paper, we propose using another physical property to eliminate noise: the phase values of ion oscillations. These values can be obtained from the Fourier transform of raw Orbitrap transient signals. Our newly developed noise filtering algorithm classifies mass spectrometry peaks into two groups based on the phase values at their maxima. Then it uses this information to calculate the appropriate intensity threshold. The algorithm was tested on mass spectra of humic substances and showed good agreement with methods previously reported in the literature. The results obtained using our algorithm indicate that signals with anomalous phase values are mostly uninformative or related to noise. Finally, we propose an algorithm for the evaluation of noise removal efficiency. It takes into account the ratio of the number of peaks with normal and anomalous phase values that will be removed at a given intensity threshold. The algorithm can be used to validate and correct noise level estimates obtained by other methods.