Machine Learning Approaches to Identify Discriminative Signatures of Volatile Organic Compounds (VOCs) from Bacteria and Fungi Using SPME-DART-MS

Abstract

Point-of-care screening tools are essential to expedite patient care and decrease reliance on slow diagnostic tools (e.g., microbial cultures) to identify pathogens and their associated antibiotic resistance. Analysis of volatile organic compounds (VOC) emitted from biological media has seen increased attention in recent years as a potential non-invasive diagnostic procedure. This work explores the use of solid phase micro-extraction (SPME) and ambient plasma ionization mass spectrometry (MS) to rapidly acquire VOC signatures of bacteria and fungi. The MS spectrum of each pathogen goes through a preprocessing and feature extraction pipeline. Various supervised and unsupervised machine learning (ML) classification algorithms are trained and evaluated on the extracted feature set. These are able to classify the type of pathogen as bacteria or fungi with high accuracy, while marked progress is also made in identifying specific strains of bacteria. This study presents a new approach for the identification of pathogens from VOC signatures collected using SPME and ambient ionization MS by training classifiers on just a few samples of data. This ambient plasma ionization and ML approach is robust, rapid, precise, and can potentially be used as a non-invasive clinical diagnostic tool for point-of-care applications.

Metabolomic and Lipidomic Profiling of Bacillus Using Two-Dimensional Tandem Mass Spectrometry

“…Lipid profiles have successfully been used to speciate bacteria using mass spectrometry − and even to detect strain level differences in antibiotic resistant bacteria . Typically, a mass spectrum can provide a metabolomic and lipidomic profile while MS/MS and collision-induced dissociation (CID) experiments provide structural information on specific features of interest. , Related methods involve the use of ion mobility coupled with MS, ,− rapid evaporative ionization mass spectrometry (REIMS), , matrix assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS), , direct analysis in real time mass spectrometry (DART-MS), , direct infusion mass spectrometry (DI-MS), and liquid chromatography coupled with mass spectrometry (LC-MS). , Many of these experiments are targeted at specific biomarkers, thus being limited to specific biological targets. Untargeted methods do not have the same limitations; however, they are time-consuming as many individual MS/MS spectra are required for each sample.…”

“…25 Typically, a mass spectrum can provide a metabolomic and lipidomic profile while MS/MS and collision-induced dissociation (CID) experiments provide structural information on specific features of interest. 24,26 Related methods involve the use of ion mobility coupled with MS, 24,26−28 rapid evaporative ionization mass spectrometry (REIMS), 29,30 matrix assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS), 31,32 direct analysis in real time mass spectrometry (DART-MS), 33,34 etry (DI-MS), 35 and liquid chromatography coupled with mass spectrometry (LC-MS). 36,37 Many of these experiments are targeted at specific biomarkers, thus being limited to specific biological targets.…”

Analytical methods for the analysis of volatile natural products

“…102 Arora et al explored the use of SPME and ambient plasma ionization mass spectrometry to rapidly acquire VOC signatures of bacteria and fungi. 103 This study presents a new approach for the identication of pathogens from VOC signatures collected using SPME and ambient ionization MS by training classiers on just a few samples of data.…”

A Mass Spectrometry-Machine Learning Approach for Detecting Volatile Organic Compound Emissions for Early Fire Detection

“…Volatile organic compound (VOC) profiling is a highly informative mass spectrometry (MS) technique that has a number of applications including disease diagnostics, chemical and environmental safety, food quality control, and others. − However, the raw output data from VOC analysis is often extensive, and data interpretation can be time-consuming and complicated. While mass spectrometry data of VOCs is typically analyzed through multivariate analysis techniques such as principal component analysis, the use of machine learning in VOC analysis has been shown to provide a more streamlined approach to processing VOC profiling data …”

“…While mass spectrometry data of VOCs is typically analyzed through multivariate analysis techniques such as principal component analysis, the use of machine learning in VOC analysis has been shown to provide a more streamlined approach to processing VOC profiling data. 4 VOC profiling has been demonstrated as a useful detection and identification tool most often in the context of microbial and disease detection. 1−4 For example, Zhu et al analyzed the volatile metabolites emitted by four species of bacteria and was able to use principal component analysis (PCA) to differentiate data from each species.…”