A novel method for micropollutant quantification using deep learning and multi-objective optimization

Daeun Yun; Daeho Kang; Jiyi Jang; Anne Therese Angeles; JongCheol Pyo; Junho Jeon; Sang-Soo Baek; Kyung Hwa Cho

doi:10.1016/j.watres.2022.118080

A novel method for micropollutant quantification using deep learning and multi-objective optimization

Water Res. 2022 Apr 1:212:118080. doi: 10.1016/j.watres.2022.118080. Epub 2022 Jan 15.

Authors

Daeun Yun¹, Daeho Kang², Jiyi Jang¹, Anne Therese Angeles¹, JongCheol Pyo³, Junho Jeon⁴, Sang-Soo Baek⁵, Kyung Hwa Cho⁶

Affiliations

¹ School of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology, 50 UNIST-gil, Ulsan 44919, South Korea.
² Department of Environmental Engineering, Changwon National University, Changwon, Gyeongsangnamdo, 51140, South Korea.
³ Center for Environmental Data Strategy, Korea Environment Institute, Sejong 30147, South Korea.
⁴ Department of Environmental Engineering, Changwon National University, Changwon, Gyeongsangnamdo, 51140, South Korea; School of Smart and Green Engineering, Changwon National University, Changwon, Gyeongsangnamdo, 51140, South Korea.
⁵ Department of Environmental Engineering, Yeungnam University, 280 Daehak-Ro, Gyeongsan-Si, Gyeongbuk 38541, South Korea. Electronic address: kbcqrkbcqr12@gmail.com.
⁶ School of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology, 50 UNIST-gil, Ulsan 44919, South Korea. Electronic address: khcho@unist.ac.kr.

PMID: 35114526
DOI: 10.1016/j.watres.2022.118080

Abstract

Micropollutants (MPs) released into aquatic ecosystems have adverse effects on public health. Hence, monitoring and managing MPs in aquatic systems are imperative. MPs can be quantified by high-resolution mass spectrometry (HRMS) with stable isotope-labeled (SIL) standards. However, high cost of SIL solutions is a significant issue. This study aims to develop a rapid and cost-effective analytical approach to estimate MP concentrations in aquatic systems based on deep learning (DL) and multi-objective optimization. We hypothesized that internal standards could quantify the MP concentrations other than the target substance. Our approach considered the precision of intra-/inter-day repeatability and natural organic matter information to reduce instrumental error and matrix effect. We selected standard solutions to estimate the concentrations of 18 MPs. Among the optimal DL models, DarkNet-53 using nine standard solutions yielded the highest performance, while ResNet-50 yielded the lowest. Overall, this study demonstrated the capability of DL models for estimating MP concentrations.

Keywords: Convolutional neural network; Deep learning; High resolution mass spectrometry; Micropollutant; Surrogate method.

MeSH terms

Deep Learning*
Ecosystem*
Isotopes
Mass Spectrometry
Reference Standards

Substances

Isotopes