Deng Fulong, He Xingliang, Yue Hanlu, Sun Hongen, Wu Bin, Zhao Zhongjun, Duan Yixiang
School of Mechanical Engineering, Sichuan University, Chengdu 610064, PR China.
Aliben Science and Technology Co., Ltd, Chengdu 610064, PR China.
J Am Soc Mass Spectrom. 2024 Dec 4;35(12):2900-2909. doi: 10.1021/jasms.4c00244. Epub 2024 Oct 1.
In the analysis of mass spectrometry, the peak identification from the overlapped region is necessary yet difficult. Although various methods have been developed to identify these peaks, especially the continuous wavelet transformation, their applications are still limited and it is hard to deal with the complex overlapped peaks. In this study, a novel peak extraction algorithm of mass spectrometry based on iterative adaptive curve fitting is proposed to address these challenges. It fully utilizes the global optimization characteristics of adaptive curve fitting. Initial peak parameters are obtained using a window searching method, and the residuals between the adaptive fitting peak and the original data indicate the fit's effectiveness and provide information about the peaks in overlap. Using this information, we performed iterative adaptive fitting, continuously updating the overlapped peaks until the residuals met the completion criteria. All of the peaks within the overlapped region can be successfully extracted by the final fitting. The proposed method is evaluated by the simulated data, the real signal from a public data set, and the spectra of two different mass spectrometry instruments. The results demonstrate that this method can more effectively extract peaks with severe overlap and multiple overlapped peaks, resist noise interference, and offer the potential to process peaks with a high dynamic range. More importantly, the proposed method accurately identifies overlapped peaks in the actual spectra from various mass spectrometry instruments, which helps the qualitative and quantitative analyses to a great extent.
在质谱分析中，从重叠区域识别峰是必要但困难的。尽管已经开发了各种方法来识别这些峰，特别是连续小波变换，但其应用仍然有限，难以处理复杂的重叠峰。在本研究中，提出了一种基于迭代自适应曲线拟合的新型质谱峰提取算法来应对这些挑战。它充分利用了自适应曲线拟合的全局优化特性。使用窗口搜索方法获得初始峰参数，自适应拟合峰与原始数据之间的残差表明拟合的有效性，并提供有关重叠峰的信息。利用这些信息，我们进行了迭代自适应拟合，不断更新重叠峰，直到残差满足完成标准。最终拟合可以成功提取重叠区域内的所有峰。通过模拟数据、来自公共数据集的真实信号以及两种不同质谱仪器的光谱对所提出的方法进行了评估。结果表明，该方法能够更有效地提取具有严重重叠和多个重叠峰的峰，抵抗噪声干扰，并具有处理高动态范围峰的潜力。更重要的是，所提出的方法能够准确识别来自各种质谱仪器的实际光谱中的重叠峰，这在很大程度上有助于定性和定量分析。
