Fusion of Quality Evaluation Metrics and Convolutional Neural Network Representations for ROI Filtering in LC–MS

Abstract

Region of interest (ROI) extraction is a fundamental step in analyzing metabolomic datasets acquired by liquid chromatography–mass spectrometry (LC–MS). However, noises and backgrounds in LC–MS data often affect the quality of extracted ROIs. Therefore, developing effective ROI evaluation algorithms is necessary to eliminate false positives meanwhile keep the false-negative rate as low as possible. In this study, a deep fused filter of ROIs (dffROI) was proposed to improve the accuracy of ROI extraction by combining the handcrafted evaluation metrics with convolutional neural network (CNN)-learned representations. To evaluate the performance of dffROI, dffROI was compared with peakonly (CNN-learned representation) and five handcrafted metrics on three LC–MS datasets and a gas chromatography–mass spectrometry (GC–MS) dataset. Results show that dffROI can achieve higher accuracy, better true-positive rate, and lower false-positive rate. Its accuracy, true-positive rate, and false-positive rate are 0.9841, 0.9869, and 0.0186 on the test set, respectively. The classification error rate of dffROI (1.59%) is significantly reduced compared with peakonly (2.73%). The model-agnostic feature importance demonstrates the necessity of fusing handcrafted evaluation metrics with the convolutional neural network representations. dffROI is an automatic, robust, and universal method for ROI filtering by virtue of information fusion and end-to-end learning. It is implemented in Python programming language and open-sourced at under BSD License. Furthermore, it has been integrated into the KPIC2 framework previously proposed by our group to facilitate real metabolomic LC–MS dataset analysis.

GCMSFormer: A Fully Automatic Method for the Resolution of Overlapping Peaks in Gas Chromatography–Mass Spectrometry

“…23,24 Deep learning is also widely used in analytical chemistry due to the availability of spectra, structures, and property databases. 25 breakthroughs in various fields of analytical chemistry, such as chromatography, 26,27 ion mobility spectrometry, 28,29 mass spectrometry, 30,31 nuclear magnetic resonance (NMR) spectroscopy, 32 Raman spectroscopy, 33−35 and infrared spectroscopy. 36,37 Specifically in the field of GC−MS, deep learning has been used for peak detection, 38 retention index prediction, 39,40 spectral library retrieval, 41 mass spectral prediction, 42,43 and overlapping peak resolution.…”

“…With the proposal of network architectures such as convolutional neural networks (CNN), graph neural networks (GNN), recurrent neural networks (RNN), and attention-based networks, it has achieved many groundbreaking advances in computer vision, natural language processing, speech recognition, and artificial intelligence (AI) for science. , Deep learning is also widely used in analytical chemistry due to the availability of spectra, structures, and property databases . It has already made breakthroughs in various fields of analytical chemistry, such as chromatography, , ion mobility spectrometry, , mass spectrometry, , nuclear magnetic resonance (NMR) spectroscopy, Raman spectroscopy, − and infrared spectroscopy. , Specifically in the field of GC–MS, deep learning has been used for peak detection, retention index prediction, , spectral library retrieval, mass spectral prediction, , and overlapping peak resolution. − In 2023, Fan et al proposed the AutoRes method based on the pseudo-Siamese convolutional neural networks (pSCNN). It can fully automate the batch processing of untargeted GC–MS data, and the entire resolution process does not require any parameters to be optimized.…”

Deep-Learning-Based Mixture Identification for Nuclear Magnetic Resonance Spectroscopy Applied to Plant Flavors

“…Recently, machine learning has witnessed significant advancements due to the availability of enhanced computational resources and novel deep learning algorithms [26,27]. These developments have enabled researchers to better deal with the challenges in chemistry, especially analytical chemistry [28], such as near infrared spectroscopy [29], Raman spectroscopy [30][31][32][33], mass spectrometry [34][35][36][37][38][39][40], chromatography [41][42][43][44][45], and ion mobility spectrometry [46][47][48]. Various machine learning methods have also been gradually applied in NMR spectroscopy [49,50], including complex mixture analysis in omics [16,51].…”