| Timetable |
| Download Conference Program |
| Download All Abstracts |
| Zoom Access |
| Corporate Program |
Poster Presentations
Day 3, June 24（Tue.）
Room P (Maesato East, Foyer, Ocean Wing)
- 3P-PM-09
Supervised Deep learning approach to automatically classify peaks of MALDI-TOF Datasets
(GRC)
oAli Farhan, Yi-Sheng Wang
Matrix-Assisted Laser Desorption Ionization mass spectrometry time of flight (MALDI-TOF) is a well-known technique in mass spectrometry that produce ions from large molecules. There are variety of software packages have been developed to analyze MALDI-TOF datasets. In current era artificial intelligence (AI) is emerging with novel machine learning (ML) and deep learning (DL) models to train the desired datasets for prediction of unknown anomaly. In this study we trained 500 laser shots data using DL model to label good, fair and bad followed by the quality of peaks in each class of dataset. Later, validation was performed on blind datasets using similar protocol for each class to fetch the data. We developed a user-friendly software package that can automatically predict the quality of TOF spectra without manual labour to set the quality parameters for peak selection. The software is designed in Python programming language that is convenient to handle with less taking memory space on computer systems as compared to MATLAB and Java. The overall accuracy of software on blind testing is 94% with f1 score 96% and precision rate 94.3%.