Mass spectral substance detections using long short-term memory networks

Junxiu Liu, J. Zhang, Y. Luo, Scott Yang, Jinling Wang, Q. Fu

Research output: Contribution to journalArticle

Abstract

In this paper, mass spectral substance detection methods are proposed, which employ long short-term memory (LSTM) recurrent neural networks to classify the mass spectrometry data and can accurately detect chemical substances. As the LSTM has the excellent understanding ability for the historical information and classification capability for the time series data, a high detection rate is obtained for the dataset which was collected by a time-of-flight proton-transfer mass spectrometer. In addition, the differential operation is used as the pre-processing method to determine the start time points of the detections which significantly improve the accuracy performance by 123%. The feature selection algorithm of Relief is also used in this paper to select the most significant channels for the mass spectrometer. It can reduce the computing resource cost, and the results show that the network size is reduced by 28% and the training speed is improved by 35%. By using these two pre-processing methods, the LSTM-based substance detection system can achieve the tradeoff between high detection rate and low computing resource consumption, which is beneficial to the devices with constraint computing resources such as low-cost embedded hardware systems.

LanguageEnglish
Article number8606092
Pages10734-10744
Number of pages11
JournalIEEE Access
Volume7
DOIs
Publication statusPublished - 9 Jan 2019

Fingerprint

Mass spectrometers
Proton transfer
Recurrent neural networks
Processing
Mass spectrometry
Feature extraction
Costs
Time series
Hardware
Long short-term memory

Keywords

  • Mass spectral substance detections
  • chemometrics
  • long short-term memory networks

Cite this

@article{2c1752101ba94e2984c2fe1462c6d9c4,
title = "Mass spectral substance detections using long short-term memory networks",
abstract = "In this paper, mass spectral substance detection methods are proposed, which employ long short-term memory (LSTM) recurrent neural networks to classify the mass spectrometry data and can accurately detect chemical substances. As the LSTM has the excellent understanding ability for the historical information and classification capability for the time series data, a high detection rate is obtained for the dataset which was collected by a time-of-flight proton-transfer mass spectrometer. In addition, the differential operation is used as the pre-processing method to determine the start time points of the detections which significantly improve the accuracy performance by 123{\%}. The feature selection algorithm of Relief is also used in this paper to select the most significant channels for the mass spectrometer. It can reduce the computing resource cost, and the results show that the network size is reduced by 28{\%} and the training speed is improved by 35{\%}. By using these two pre-processing methods, the LSTM-based substance detection system can achieve the tradeoff between high detection rate and low computing resource consumption, which is beneficial to the devices with constraint computing resources such as low-cost embedded hardware systems.",
keywords = "Mass spectral substance detections, chemometrics, long short-term memory networks",
author = "Junxiu Liu and J. Zhang and Y. Luo and Scott Yang and Jinling Wang and Q. Fu",
note = "Accepted dated listed on the uploaded article",
year = "2019",
month = "1",
day = "9",
doi = "10.1109/ACCESS.2019.2891548",
language = "English",
volume = "7",
pages = "10734--10744",
journal = "IEEE Access",
issn = "2169-3536",

}

Mass spectral substance detections using long short-term memory networks. / Liu, Junxiu; Zhang, J.; Luo, Y.; Yang, Scott; Wang, Jinling; Fu, Q.

In: IEEE Access, Vol. 7, 8606092, 09.01.2019, p. 10734-10744.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Mass spectral substance detections using long short-term memory networks

AU - Liu, Junxiu

AU - Zhang, J.

AU - Luo, Y.

AU - Yang, Scott

AU - Wang, Jinling

AU - Fu, Q.

N1 - Accepted dated listed on the uploaded article

PY - 2019/1/9

Y1 - 2019/1/9

N2 - In this paper, mass spectral substance detection methods are proposed, which employ long short-term memory (LSTM) recurrent neural networks to classify the mass spectrometry data and can accurately detect chemical substances. As the LSTM has the excellent understanding ability for the historical information and classification capability for the time series data, a high detection rate is obtained for the dataset which was collected by a time-of-flight proton-transfer mass spectrometer. In addition, the differential operation is used as the pre-processing method to determine the start time points of the detections which significantly improve the accuracy performance by 123%. The feature selection algorithm of Relief is also used in this paper to select the most significant channels for the mass spectrometer. It can reduce the computing resource cost, and the results show that the network size is reduced by 28% and the training speed is improved by 35%. By using these two pre-processing methods, the LSTM-based substance detection system can achieve the tradeoff between high detection rate and low computing resource consumption, which is beneficial to the devices with constraint computing resources such as low-cost embedded hardware systems.

AB - In this paper, mass spectral substance detection methods are proposed, which employ long short-term memory (LSTM) recurrent neural networks to classify the mass spectrometry data and can accurately detect chemical substances. As the LSTM has the excellent understanding ability for the historical information and classification capability for the time series data, a high detection rate is obtained for the dataset which was collected by a time-of-flight proton-transfer mass spectrometer. In addition, the differential operation is used as the pre-processing method to determine the start time points of the detections which significantly improve the accuracy performance by 123%. The feature selection algorithm of Relief is also used in this paper to select the most significant channels for the mass spectrometer. It can reduce the computing resource cost, and the results show that the network size is reduced by 28% and the training speed is improved by 35%. By using these two pre-processing methods, the LSTM-based substance detection system can achieve the tradeoff between high detection rate and low computing resource consumption, which is beneficial to the devices with constraint computing resources such as low-cost embedded hardware systems.

KW - Mass spectral substance detections

KW - chemometrics

KW - long short-term memory networks

UR - http://www.scopus.com/inward/record.url?scp=85061188995&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2019.2891548

DO - 10.1109/ACCESS.2019.2891548

M3 - Article

VL - 7

SP - 10734

EP - 10744

JO - IEEE Access

T2 - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 8606092

ER -