Vol.11 (2021) No. 4 ISSN: 2088-5334 Rapid Detection of Escherchia coli and Salmonella Typhimurium Using Lab-Made Electronic Nose Coupled with Chemometric Tools Wredha Sandhi Ardha Prakoso a,b , Prima Febri Astantri a,b , Kuwat Triyana c,f , Tri Untari d , Claude Mona Airin e, Pudji Astuti e,* a Postgraduate of Veterinary Science, Faculty of Veterinary Medicine, Gadjah Mada University, Yogyakarta 55281, Indonesia b Animal Quarantine, Agriculture Quarantine Agency, Ministry of Agriculture, Jakarta 12550, Indonesia c Department of Physics, Faculty of Mathematics and Natural Science, Gadjah Mada University, Yogyakarta 55281, Indonesia d Department of Microbiology, Faculty of Veterinary Medicine, Gadjah Mada University, Yogyakarta 55281, Indonesia e Department of Physiology, Faculty of Veterinary Medicine, Gadjah Mada University, Yogyakarta 55281, Indonesia f Institute of Halal Industry and System (IHIS), Gadjah Mada University, Yogyakarta 55281, Indonesia Corresponding author: * pastuti2@ugm.ac.id Abstract—This study aims to investigate the performance of a lab-made electronic nose coupled with chemometric tools for detecting Escherichia coli (E. coli) and Salmonella Typhimurium (S. Typhimurium) inoculated in media. The pathogenic E. coli and S. Typhimurium play a significant role as the agent causing food-borne diseases, posing a threat to human health worldwide. Some advanced analytical instruments like RT-PCR and GC-MS are often used for detecting such pathogenic bacteria. Unfortunately, they are not suitable for rapid and routine measurements because of time-consuming, require experts, and complicated sample preparation. Otherwise, electronic nose (e-nose) has been reported to be successful for profiling volatile compounds released by various biological materials. The e-nose comprised eight types of metal oxide gas sensors connected with a data acquisition system and chemometric tools. For this purpose, Fast Fourier Transform (FFT) was applied for signal pre-processing and feature extraction to all datasets collected by the sensor array in the e-nose. Furthermore, chemometric tools are used for classification models of all extracted features, including linear and quadratic discriminant analysis (LDA and QDA) and support vector machine (SVM). As a result, SVM showed the highest performance, enabling identifying E. coli and S. Typhimurium inoculated TSB with an accuracy of 99% and 98%, respectively. Among the chemometric tools, the e-nose-SVM also resulted in the highest accuracy in differentiating E. coli from S. Typhimurium of 84%. These results motivated e-nose to have a high prospect to rapidly detect such bacteria for food safety and quality control inspection, particularly potential quarantine products. Keywords— Electronic nose; Escherichia coli; Salmonella Typhimurium; chemometric; metabolic volatile organic compounds. Manuscript received 7 Aug. 2020; revised 3 Feb. 2021; accepted 20 Feb. 2021. Date of publication 31 Aug. 2021. IJASEIT is licensed under a Creative Commons Attribution-Share Alike 4.0 International License. I. INTRODUCTION Contaminated food by pathogenic bacteria has been posing a crucial issue due to the threat to human health and life worldwide. Millions of people suffered, and thousands died every year due to these food-borne disease outbreaks, even though certain bacteria such as Bacillus subtilis natto can degrade insoluble fibrin fibers of thrombosis [1]. Food is one of the media that facilitates the growth of pathogenic microorganisms to become an agent of food-borne diseases. More than 90% of food contamination cases in developing countries are caused by bacteria, such as S. Typhimurium and E. coli [2]–[4]. E. coli produces verotoxigenic, the primary source of food-borne diseases in humans. Meanwhile, S. Typhimurium is classified as a nontyphoidal serovar Salmonella (NTS) which can cause fever, nausea, cramps, headaches, diarrhea, and vomiting in humans [3], [5]. Also, food changes its taste and aroma when it is contaminated with bacterial like E. coli, decreasing its quality [6]. Contamination of Salmonella sp. in food is prohibited by the Indonesian standard of food safety (SNI), while E. coli contamination in food is allowed up to a maximum of 10 CFU/g. Conventional methods such as bacterial count calculations, growing on selective media, and simple biochemistry analysis methods are usually used to identify 1494