abs

학술발표회초록보기

Inquiry on Abstract abstract@kcsnet.or.kr

Inquiry on Payment member@kcsnet.or.kr

현재 가능한 작업은 아래와 같습니다.
  • 03월 02일 17시 이후 : 초록수정 불가능, 일정확인 및 검색만 가능

129th General Meeting of Korean Chemical Society & Exposition AutoMRM: a targeted proteomics data interpretation tool based on convolutional neural networks and explainable artificial intelligence

Submission Date :
2 / 28 / 2022 , 10 : 22 : 48
Abstract Number :
129022830668
Presenting Type:
Poster Presentation Analytical Chemistry Oral Presentation
Presenting Area :
Analytical Chemistry
Authors :
Jungkap Park, Jiwon Hong1, Seunghoon Back1, Hokeun Kim1, Sang-Won Lee1, Sangtae Kim2,*
Bioinformatics Group, Bertis Inc., Korea
1Department of Chemistry, Korea University, Korea
2Bioinformatics Group, Bertis Bioscience Inc., United States
※ You will get a confirmation e-mail on approval or disapproval of the author's
overseas affiliation in a few days.
Approved : 1 case
Assigned Code :
ANAL.P-356 Assigend Code Guideline
Presenting Time :
April 14 (THU) 11:00~13:00

Deep neural networks have led to breakthroughs in discovery proteomics, but their adoption in targeted proteomics has been slow. In clinical proteomics laboratories, researchers spend a significant time on manual peak picking, interference identification, and peak area adjustments to interpret multiple reaction monitoring (MRM) or parallel reaction monitoring (PRM) data. The burden of manual inspection is a major factor limiting transferability, reproducibility, and scalability of targeted proteomics in clinical applications. We present AutoMRM, a targeted proteomics data interpretation tool based on convolutional neural networks (CNN) and explainable artificial intelligence (XAI), designed for clinical mass spectrometry laboratories. When applied to MRM data, AutoMRM shows an accuracy comparable to that of human experts, obviating or significantly reducing the burden of manual inspection. With AutoMRM, an MRM analysis task which used to take over 600 hours by human experts could be completed in less than 5 minutes. We plan to apply the method to the PRM and data-independent acquisition data.