Toward an Integrated Machine Learning Model of a Proteomics Experiment

Benjamin A. Neely; Viktoria Dorfer; Lennart Martens; Isabell Bludau; Robbin Bouwmeester; Sven Degroeve; Eric W. Deutsch; Siegfried Gessulat; Lukas Käll; Pawel Palczynski; Samuel H. Payne; Tobias Greisager Rehfeldt; Tobias Schmidt; Veit Schwämmle; Julian Uszkoreit; Juan Antonio Vizcaíno; Mathias Wilhelm; Magnus Palmblad

doi:10.1021/acs.jproteome.2c00711

Toward an Integrated Machine Learning Model of a Proteomics Experiment

Benjamin A. Neely^*
, Viktoria Dorfer^*
, Lennart Martens^*
, Isabell Bludau
, Robbin Bouwmeester
, Sven Degroeve
, Eric W. Deutsch
, Siegfried Gessulat
, Lukas Käll
, Pawel Palczynski
, Samuel H. Payne
, Tobias Greisager Rehfeldt
, Tobias Schmidt
, Veit Schwämmle
, Julian Uszkoreit
, Juan Antonio Vizcaíno
, Mathias Wilhelm
, Magnus Palmblad^*

^*Corresponding author for this work

National Institute of Standards and Technology
University of Applied Sciences Upper Austria
Max Planck Institute of Biochemistry
Institute for Systems Biology
MSAID Gmbh
KTH Royal Institute of Technology
Brigham Young University
Ruhr University Bochum
Technical University of Munich
Leiden University
Ghent University
Wellcome Trust Genome Campus

Research output: Contribution to journal › Journal article › Research › peer-review

198 Downloads (Pure)

Abstract

In recent years machine learning has made extensive progress in modeling many aspects of mass spectrometry data. We brought together proteomics data generators, repository managers, and machine learning experts in a workshop with the goals to evaluate and explore machine learning applications for realistic modeling of data from multidimensional mass spectrometry-based proteomics analysis of any sample or organism. Following this sample-To-data roadmap helped identify knowledge gaps and define needs. Being able to generate bespoke and realistic synthetic data has legitimate and important uses in system suitability, method development, and algorithm benchmarking, while also posing critical ethical questions. The interdisciplinary nature of the workshop informed discussions of what is currently possible and future opportunities and challenges. In the following perspective we summarize these discussions in the hope of conveying our excitement about the potential of machine learning in proteomics and to inspire future research.

Original language	English
Journal	Journal of Proteome Research
Volume	22
Issue number	3
Pages (from-to)	681–696
ISSN	1535-3893
DOIs	https://doi.org/10.1021/acs.jproteome.2c00711
Publication status	Published - 6. Feb 2023

Bibliographical note

Funding Information:
The Lorentz Center workshop was funded by the Dutch Research Council (NWO) with generous support from the Leiden University Medical Center, Thermo Fisher Scientific, and the Journal of Proteome Research that made this workshop possible. The authors thank Dr. Mike Preuss and Dr. Elisabeth Bik for their inspiring lectures and active participation in discussions on machine learning and research integrity during the workshop. Special thanks go to all the Lorentz Center staff who went above and beyond their duties to make this hybrid workshop a success in COVID times. Identification of certain commercial equipment, instruments, software, or materials does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the products identified are necessarily the best available for the purpose.

Funding

The Lorentz Center workshop was funded by the Dutch Research Council (NWO) with generous support from the Leiden University Medical Center, Thermo Fisher Scientific, and the Journal of Proteome Research that made this workshop possible. The authors thank Dr. Mike Preuss and Dr. Elisabeth Bik for their inspiring lectures and active participation in discussions on machine learning and research integrity during the workshop. Special thanks go to all the Lorentz Center staff who went above and beyond their duties to make this hybrid workshop a success in COVID times. Identification of certain commercial equipment, instruments, software, or materials does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the products identified are necessarily the best available for the purpose.

Keywords

artificial intelligence
deep learning
enzymatic digestion
ion mobility
liquid chromatography
machine learning
research integrity
synthetic data
tandem mass spectrometry

Documents & Links

10.1021/acs.jproteome.2c00711Licence: CC BY

Open access versionFinal published version, 1.9 MBLicence: CC BY

SDU Link

Check access to the material in SDU Library's search engine

1 Ph.D. thesis

Application of deep learning methods on publicly available mass spectrometry-based proteomics data
Rehfeldt, T. G., 4. Nov 2023, Syddansk Universitet. Det Naturvidenskabelige Fakultet. 155 p.
Research output: Thesis › Ph.D. thesis

Open Access
File
212 Downloads (Pure)

Cite this

Neely, B. A., Dorfer, V., Martens, L., Bludau, I., Bouwmeester, R., Degroeve, S., Deutsch, E. W., Gessulat, S., Käll, L., Palczynski, P., Payne, S. H., Rehfeldt, T. G., Schmidt, T., Schwämmle, V., Uszkoreit, J., Vizcaíno, J. A., Wilhelm, M., & Palmblad, M. (2023). Toward an Integrated Machine Learning Model of a Proteomics Experiment. Journal of Proteome Research, 22(3), 681–696. https://doi.org/10.1021/acs.jproteome.2c00711

@article{51e5cb6e1482401c99be7304311b254c,

title = "Toward an Integrated Machine Learning Model of a Proteomics Experiment",

abstract = "In recent years machine learning has made extensive progress in modeling many aspects of mass spectrometry data. We brought together proteomics data generators, repository managers, and machine learning experts in a workshop with the goals to evaluate and explore machine learning applications for realistic modeling of data from multidimensional mass spectrometry-based proteomics analysis of any sample or organism. Following this sample-To-data roadmap helped identify knowledge gaps and define needs. Being able to generate bespoke and realistic synthetic data has legitimate and important uses in system suitability, method development, and algorithm benchmarking, while also posing critical ethical questions. The interdisciplinary nature of the workshop informed discussions of what is currently possible and future opportunities and challenges. In the following perspective we summarize these discussions in the hope of conveying our excitement about the potential of machine learning in proteomics and to inspire future research.",

keywords = "artificial intelligence, deep learning, enzymatic digestion, ion mobility, liquid chromatography, machine learning, research integrity, synthetic data, tandem mass spectrometry",

author = "Neely, \{Benjamin A.\} and Viktoria Dorfer and Lennart Martens and Isabell Bludau and Robbin Bouwmeester and Sven Degroeve and Deutsch, \{Eric W.\} and Siegfried Gessulat and Lukas K{\"a}ll and Pawel Palczynski and Payne, \{Samuel H.\} and Rehfeldt, \{Tobias Greisager\} and Tobias Schmidt and Veit Schw{\"a}mmle and Julian Uszkoreit and Vizca{\'i}no, \{Juan Antonio\} and Mathias Wilhelm and Magnus Palmblad",

note = "Funding Information: The Lorentz Center workshop was funded by the Dutch Research Council (NWO) with generous support from the Leiden University Medical Center, Thermo Fisher Scientific, and the Journal of Proteome Research that made this workshop possible. The authors thank Dr. Mike Preuss and Dr. Elisabeth Bik for their inspiring lectures and active participation in discussions on machine learning and research integrity during the workshop. Special thanks go to all the Lorentz Center staff who went above and beyond their duties to make this hybrid workshop a success in COVID times. Identification of certain commercial equipment, instruments, software, or materials does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the products identified are necessarily the best available for the purpose.",

year = "2023",

month = feb,

day = "6",

doi = "10.1021/acs.jproteome.2c00711",

language = "English",

volume = "22",

pages = "681{\textendash}696",

journal = "Journal of Proteome Research",

issn = "1535-3893",

publisher = "American Chemical Society",

number = "3",

}

Neely, BA, Dorfer, V, Martens, L, Bludau, I, Bouwmeester, R, Degroeve, S, Deutsch, EW, Gessulat, S, Käll, L, Palczynski, P, Payne, SH, Rehfeldt, TG, Schmidt, T, Schwämmle, V, Uszkoreit, J, Vizcaíno, JA, Wilhelm, M & Palmblad, M 2023, 'Toward an Integrated Machine Learning Model of a Proteomics Experiment', Journal of Proteome Research, vol. 22, no. 3, pp. 681–696. https://doi.org/10.1021/acs.jproteome.2c00711

TY - JOUR

T1 - Toward an Integrated Machine Learning Model of a Proteomics Experiment

AU - Neely, Benjamin A.

AU - Dorfer, Viktoria

AU - Martens, Lennart

AU - Bludau, Isabell

AU - Bouwmeester, Robbin

AU - Degroeve, Sven

AU - Deutsch, Eric W.

AU - Gessulat, Siegfried

AU - Käll, Lukas

AU - Palczynski, Pawel

AU - Payne, Samuel H.

AU - Rehfeldt, Tobias Greisager

AU - Schmidt, Tobias

AU - Schwämmle, Veit

AU - Uszkoreit, Julian

AU - Vizcaíno, Juan Antonio

AU - Wilhelm, Mathias

AU - Palmblad, Magnus

N1 - Funding Information: The Lorentz Center workshop was funded by the Dutch Research Council (NWO) with generous support from the Leiden University Medical Center, Thermo Fisher Scientific, and the Journal of Proteome Research that made this workshop possible. The authors thank Dr. Mike Preuss and Dr. Elisabeth Bik for their inspiring lectures and active participation in discussions on machine learning and research integrity during the workshop. Special thanks go to all the Lorentz Center staff who went above and beyond their duties to make this hybrid workshop a success in COVID times. Identification of certain commercial equipment, instruments, software, or materials does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the products identified are necessarily the best available for the purpose.

PY - 2023/2/6

Y1 - 2023/2/6

N2 - In recent years machine learning has made extensive progress in modeling many aspects of mass spectrometry data. We brought together proteomics data generators, repository managers, and machine learning experts in a workshop with the goals to evaluate and explore machine learning applications for realistic modeling of data from multidimensional mass spectrometry-based proteomics analysis of any sample or organism. Following this sample-To-data roadmap helped identify knowledge gaps and define needs. Being able to generate bespoke and realistic synthetic data has legitimate and important uses in system suitability, method development, and algorithm benchmarking, while also posing critical ethical questions. The interdisciplinary nature of the workshop informed discussions of what is currently possible and future opportunities and challenges. In the following perspective we summarize these discussions in the hope of conveying our excitement about the potential of machine learning in proteomics and to inspire future research.

AB - In recent years machine learning has made extensive progress in modeling many aspects of mass spectrometry data. We brought together proteomics data generators, repository managers, and machine learning experts in a workshop with the goals to evaluate and explore machine learning applications for realistic modeling of data from multidimensional mass spectrometry-based proteomics analysis of any sample or organism. Following this sample-To-data roadmap helped identify knowledge gaps and define needs. Being able to generate bespoke and realistic synthetic data has legitimate and important uses in system suitability, method development, and algorithm benchmarking, while also posing critical ethical questions. The interdisciplinary nature of the workshop informed discussions of what is currently possible and future opportunities and challenges. In the following perspective we summarize these discussions in the hope of conveying our excitement about the potential of machine learning in proteomics and to inspire future research.

KW - artificial intelligence

KW - deep learning

KW - enzymatic digestion

KW - ion mobility

KW - liquid chromatography

KW - machine learning

KW - research integrity

KW - synthetic data

KW - tandem mass spectrometry

U2 - 10.1021/acs.jproteome.2c00711

DO - 10.1021/acs.jproteome.2c00711

M3 - Journal article

C2 - 36744821

AN - SCOPUS:85147873620

SN - 1535-3893

VL - 22

SP - 681

EP - 696

JO - Journal of Proteome Research

JF - Journal of Proteome Research

IS - 3

ER -

Toward an Integrated Machine Learning Model of a Proteomics Experiment

Abstract

Bibliographical note

Funding

Keywords

Documents & Links

SDU Link

Fingerprint

Related research output

Application of deep learning methods on publicly available mass spectrometry-based proteomics data

Cite this