Deconvolution of overlapping peaks in ion mobility spectrometry based on a multiobjective dynamic teaching‐learning‐based optimization

Abstract

Rationale Because of its powerful analytical ability, ion mobility spectrometry (IMS) plays an important role in the field of mass spectrometry. However, one of the main defects of IMS is its low structural resolution, which leads to the phenomenon of peak overlap in the analysis of compounds with similar mass charge ratio. Methods A multiobjective dynamic teaching‐learning‐based optimization (MDTLBO) method was proposed to separate IMS overlapping peaks. This method prevents local optimization and identifies peak model coefficients efficiently. In addition, the position information of particles largely reflects the half‐peak width of IMS, which makes single peaks difficult to appear and coefficient identification easier. Results The performance comparison of MDTLBO with other deconvolution methods (genetic algorithm, improved particle swarm optimization algorithm, and dynamic inertia weight particle swarm optimization algorithm) shows that the maximum deconvolution error of MDTLBO is only 0.7%, which is much lower than that for the other three methods. In addition, robustness is a performance index that reflects the advantages and disadvantages of the algorithm. Conclusion MBTLBO is more robust than other algorithms for separating overlapping peaks. The algorithm can separate the heavily overlapped mobility peaks, produce better analysis results, and improve the resolution of IMS.

Radial basis function neural network optimization algorithm based on dynamic inertial weight particle swarm optimization for separating overlapping peaks in ion mobility spectrometry

“…The key contributions of this research are as follows: (1) introducing the novel DIWPSO‐RBF algorithm, which combines the dynamic inertia weight particle swarm and radial basis neural network concepts; (2) enhancing the exploration and development of the optimization search space using the proposed DIWPSO‐RBF algorithm; and (3) successfully applying the DIWPSO‐RBF method to separate overlapping peaks in ion mobility spectra of two pairs of carbohydrate isomers (3′/6′‐sialyllactose [3′/6′‐SL]; fructose‐6‐phosphate [F6P], glucose‐1‐phosphate [G1P], and glucose‐6‐phosphate [G6P]). Comparative experiments against three other optimization algorithms (SSA, DIWPSO, and multi‐objective dynamic teaching‐learning‐based optimization [MDTLBO] 29 ) validate the effectiveness of the proposed algorithm, demonstrating superior search accuracy, stability, and avoidance of local optima.…”

Overlapping peak separation algorithm for ion mobility spectra based on multistrategy JAYA

“…However, this method's inertia weight size is determined by the population's average adaptive degree, leading to weakened particle diversity and reduced global exploration ability. MDTLBO has better robustness and automatically identifies its components even in high overlapping peaks 16 . However, the MDTLBO algorithm does not inherit the advantages of TLBO without special parameters and introduces dynamic factors $$$ \sigma $$$ that increase the algorithm's usage difficulty, limiting its application.…”

“…MDTLBO has better robustness and automatically identifies its components even in high overlapping peaks. 16 However, the MDTLBO algorithm does not inherit the advantages of TLBO without special parameters and introduces dynamic factors σ that increase the algorithm's usage difficulty, limiting its application. Every intelligent algorithm requires parameters such as population size and the number of iterations, which are generally referred to as common parameters.…”