Using infrared gas sensors in an in-vitro dynamic gut model for detecting short-chain fatty-acids: Technical Report

TL;DR

This paper presents a novel infrared sensor system combined with machine learning for real-time detection of short-chain fatty acids in an in-vitro gut model, aiming to improve understanding of gut health and enable rapid bioreactor monitoring.

Contribution

It introduces a new infrared sensor technology integrated with machine learning for online detection of SCFAs in bioreactors, enhancing measurement accuracy and operational efficiency.

Findings

Successful detection of SCFAs using the infrared sensor

Effective machine learning algorithms processed sensor signals

Potential for real-time monitoring and quality control in gut models

Abstract

Short-chain fatty acids (SCFAs), including acetate, propionate and butyrate, are organic fatty-acids that are produced when indigested carbohydrates are fermented in the colon by gut-bacteria. Butyrate is especially considered a beneficial compound in relation to gut health and the maintenance of colonic homeostasis. Little is known about butyrate production and the current measurement methods of fecal samples are not representative and there is a strong unmet need to measure SCFAs in-vivo to further understand their role and effect on colonic function in human gut health and disease. The general aim of the project is to develop a novel sensor based on infrared technology that is capable of online detecting short chain fatty acids (SCFA; e.g. acetate, propionate and butyrate) and other small biomolecules (e.g. ammonium) produced in bioreactors. Whereas currently such levels of small biomolecules are usually not detected online due to practical constraints (sampling-analysis-processing of data), reliable online detection could significantly decrease analysis costs. Further, it would allow to perform online quality control of the bioreactors and allow to e.g. detect deviating values at an early stage. Hence, one can react much faster tackling certain problems which will further improve the quality of each experiment. Using machine learning algorithms to process the sensor signals, the results from testing in a so-called artificial gut setup are described in this technical report.