Prosit - Deep learning model for proteomics

Prosit (https://www.proteomicsdb.org/prosit/) is a free to use deep learning framework for generating custom spectral libraries and iRT prediction and inferrence of high quality MS2 spectra for any organism and protease. Prosit is part of the ProteomeTools project and was trained on the project's high quality synthetic datasets.

Prosit has a wide range of functions, such as predicting spectra for proteases other than trypsin, generating spectral libraries for data-independent acquisition (DIA) and improving the analysis of metaproteomes. Prosit is integrated into ProteomicsDB, allowing search result re-scoring and custom spectral library generation for any organism on the basis of peptide sequence alone.

Available theses and research projects

Interested? Do you want to contribute? Are you looking for a research project, thesis, internship or simply a diversion? Visit our Open Research Projects and find topics related to [Prosit] or other projects.

Relevant source code

Using Prosit:

oktoberfest - a python package to perform rescoring, spectral library generation and collision energy calibration
koina - a pythong package and publicly available endpoint to retrieve predictions from various models (incl. Prosit) in real-time
prosit_grpc - a python client to retrieve predictions in real-time. This package is deprecated since the development of koina

Datasets:

PROSPECT - a large annotated dataset leveraging the raw data from ProteomeTools

Training Prosit:

DLOmix - python framework to train deep learning models in proteomics research

Utilities:

spectrum_fundamentals - a python library for basic processing and annotation tasks specifically for generating training/testing/holdout data
spectrum_io - a pythong package for basic spectrum I/O opperations

Primary literature

SucheHinweis: Die „Schnellsuche“ findet nur Text in den angezeigten Feldern; nicht in Abstracts oder Schlagwörtern. Der Suchbegriff muss mindestens 3 Buchstaben lang sein.

2024

Adams, Charlotte; Gabriel, Wassim; Laukens, Kris; Picciani, Mario; Wilhelm, Mathias; Bittremieux, Wout; Boonen, Kurt: Fragment ion intensity prediction improves the identification rate of non-tryptic peptides in timsTOF. Nature Communications 15 (1), 2024 mehr…
Gabriel, Wassim; Picciani, Mario; The, Matthew; Wilhelm, Mathias: Deep Learning-Assisted Analysis of Immunopeptidomics Data. In: Methods in Molecular Biology. Springer US, 2024 mehr…
Kalhor, Mostafa; Lapin, Joel; Picciani, Mario; Wilhelm, Mathias: Rescoring Peptide Spectrum Matches: Boosting Proteomics Performance by Integrating Peptide Property Predictors into Peptide Identification. Molecular & Cellular Proteomics, 2024, 100798 mehr…
Klaproth-Andrade, Daniela; Hingerl, Johannes; Bruns, Yanik; Smith, Nicholas H.; Träuble, Jakob; Wilhelm, Mathias; Gagneur, Julien: Deep learning-driven fragment ion series classification enables highly precise and sensitive de novo peptide sequencing. Nature Communications 15 (1), 2024 mehr…
The, Matthew; Picciani, Mario; Jensen, Cecilia; Gabriel, Wassim; Kuster, Bernhard; Wilhelm, Mathias: AI-Assisted Processing Pipeline to Boost Protein Isoform Detection. In: Methods in Molecular Biology. Springer US, 2024 mehr…

2023

Chiva, Cristina; Elhamraoui, Zahra; Solé, Amanda; Serret, Marc; Wilhelm, Mathias; Sabidó, Eduard: Assessment and Prediction of Human Proteotypic Peptide Stability for Proteomics Quantification. Analytical Chemistry 95 (37), 2023, 13746-13749 mehr…
McGann, Chris D.; Barshop, William D.; Canterbury, Jesse D.; Lin, Chuwei; Gabriel, Wassim; Huang, Jingjing; Bergen, David; Zabrouskov, Vlad; Melani, Rafael D.; Wilhelm, Mathias; McAlister, Graeme C.; Schweppe, Devin K.: Real-Time Spectral Library Matching for Sample Multiplexed Quantitative Proteomics. Journal of Proteome Research 22 (9), 2023, 2836-2846 mehr…
Neely, Benjamin A.; Dorfer, Viktoria; Martens, Lennart; Bludau, Isabell; Bouwmeester, Robbin; Degroeve, Sven; Deutsch, Eric W.; Gessulat, Siegfried; Käll, Lukas; Palczynski, Pawel; Payne, Samuel H.; Rehfeldt, Tobias Greisager; Schmidt, Tobias; Schwämmle, Veit; Uszkoreit, Julian; Vizcaíno, Juan Antonio; Wilhelm, Mathias; Palmblad, Magnus: Toward an Integrated Machine Learning Model of a Proteomics Experiment. Journal of Proteome Research 22 (3), 2023, 681-696 mehr…
Picciani, Mario; Gabriel, Wassim; Giurcoiu, Victor‐George; Shouman, Omar; Hamood, Firas; Lautenbacher, Ludwig; Jensen, Cecilia Bang; Müller, Julian; Kalhor, Mostafa; Soleymaniniya, Armin; Kuster, Bernhard; The, Matthew; Wilhelm, Mathias: Oktoberfest: Open‐source spectral library generation and rescoring pipeline based on Prosit. PROTEOMICS 24 (8), 2023 mehr…
Tretter, Celina; de Andrade Krätzig, Niklas; Pecoraro, Matteo; Lange, Sebastian; Seifert, Philipp; von Frankenberg, Clara; Untch, Johannes; Zuleger, Gabriela; Wilhelm, Mathias; Zolg, Daniel P.; Dreyer, Florian S.; Bräunlein, Eva; Engleitner, Thomas; Uhrig, Sebastian; Boxberg, Melanie; Steiger, Katja; Slotta-Huspenina, Julia; Ochsenreither, Sebastian; von Bubnoff, Nikolas; Bauer, Sebastian; Boerries, Melanie; Jost, Philipp J.; Schenck, Kristina; Dresing, Iska; Bassermann, Florian; Friess, Helmut; Reim, Daniel; Grützmann, Konrad; Pfütze, Katrin; Klink, Barbara; Schröck, Evelin; Haller, Bernhard; Kuster, Bernhard; Mann, Matthias; Weichert, Wilko; Fröhling, Stefan; Rad, Roland; Hiltensperger, Michael; Krackhardt, Angela M.: Proteogenomic analysis reveals RNA as a source for tumor-agnostic neoantigen identification. Nature Communications 14 (1), 2023 mehr…

2022

Ekvall, Markus; Truong, Patrick; Gabriel, Wassim; Wilhelm, Mathias; Käll, Lukas: Prosit Transformer: A transformer for Prediction of MS2 Spectrum Intensities. Journal of Proteome Research 21 (5), 2022, 1359-1364 mehr…
Gabriel, Wassim; The, Matthew; Zolg, Daniel P.; Bayer, Florian P.; Shouman, Omar; Lautenbacher, Ludwig; Schnatbaum, Karsten; Zerweck, Johannes; Knaute, Tobias; Delanghe, Bernard; Huhmer, Andreas; Wenschuh, Holger; Reimer, Ulf; Médard, Guillaume; Kuster, Bernhard; Wilhelm, Mathias: Prosit-TMT: Deep Learning Boosts Identification of TMT-Labeled Peptides. Analytical Chemistry, 2022 mehr…
Giansanti, Piero; Samaras, Patroklos; Bian, Yangyang; Meng, Chen; Coluccio, Andrea; Frejno, Martin; Jakubowsky, Hannah; Dobiasch, Sophie; Hazarika, Rashmi R.; Rechenberger, Julia; Calzada-Wack, Julia; Krumm, Johannes; Mueller, Sebastian; Lee, Chien-Yun; Wimberger, Nicole; Lautenbacher, Ludwig; Hassan, Zonera; Chang, Yun-Chien; Falcomatà, Chiara; Bayer, Florian P.; Bärthel, Stefanie; Schmidt, Tobias; Rad, Roland; Combs, Stephanie E.; The, Matthew; Johannes, Frank; Saur, Dieter; de Angelis, Martin Hrabe; Wilhelm, Mathias; Schneider, Günter; Kuster, Bernhard: Mass spectrometry-based draft of the mouse proteome. Nature Methods, 2022 mehr…
Lautenbacher, Ludwig; Samaras, Patroklos; Muller, Julian; Grafberger, Andreas; Shraideh, Marwin; Rank, Johannes; Fuchs, Simon T; Schmidt, Tobias K; The, Matthew; Dallago, Christian; Wittges, Holger; Rost, Burkhard; Krcmar, Helmut; Kuster, Bernhard; Wilhelm, Mathias: ProteomicsDB: toward a FAIR open-source resource for life-science research. Nucleic Acids Research, 2022 mehr…

2021

Schmidt, Tobias; Samaras, Patroklos; Dorfer, Viktoria; Panse, Christian; Kockmann, Tobias; Bichmann, Leon; van Puyvelde, Bart; Perez-Riverol, Yasset; Deutsch, Eric W.; Kuster, Bernhard; Wilhelm, Mathias: Universal Spectrum Explorer: A Standalone (Web-)Application for Cross-Resource Spectrum Comparison. Journal of Proteome Research 20 (6), 2021, 3388-3394 mehr…
Verbruggen, Steven; Gessulat, Siegfried; Gabriels, Ralf; Matsaroki, Anna; Van de Voorde, Hendrik; Kuster, Bernhard; Degroeve, Sven; Martens, Lennart; Van Criekinge, Wim; Wilhelm, Mathias; Menschaert, Gerben: Spectral Prediction Features as a Solution for the Search Space Size Problem in Proteogenomics. Molecular & Cellular Proteomics 20, 2021, 100076 mehr…
Wilhelm, Mathias; Zolg, Daniel P.; Graber, Michael; Gessulat, Siegfried; Schmidt, Tobias; Schnatbaum, Karsten; Schwencke-Westphal, Celina; Seifert, Philipp; de Andrade Krätzig, Niklas; Zerweck, Johannes; Knaute, Tobias; Bräunlein, Eva; Samaras, Patroklos; Lautenbacher, Ludwig; Klaeger, Susan; Wenschuh, Holger; Rad, Roland; Delanghe, Bernard; Huhmer, Andreas; Carr, Steven A.; Clauser, Karl R.; Krackhardt, Angela M.; Reimer, Ulf; Kuster, Bernhard: Deep learning boosts sensitivity of mass spectrometry-based immunopeptidomics. Nature Communications 12 (1), 2021 mehr…

2020

Searle, Brian C.; Swearingen, Kristian E.; Barnes, Christopher A.; Schmidt, Tobias; Gessulat, Siegfried; Küster, Bernhard; Wilhelm, Mathias: Generating high quality libraries for DIA MS with empirically corrected peptide predictions. Nature Communications 11 (1), 2020 mehr…

2019

Gessulat, Siegfried; Schmidt, Tobias; Zolg, Daniel Paul; Samaras, Patroklos; Schnatbaum, Karsten; Zerweck, Johannes; Knaute, Tobias; Rechenberger, Julia; Delanghe, Bernard; Huhmer, Andreas; Reimer, Ulf; Ehrlich, Hans-Christian; Aiche, Stephan; Kuster, Bernhard; Wilhelm, Mathias: Prosit: proteome-wide prediction of peptide tandem mass spectra by deep learning. Nature Methods 16 (6), 2019, 509-518 mehr…
Verbruggen, Steven; Ndah, Elvis; Van Criekinge, Wim; Gessulat, Siegfried; Kuster, Bernhard; Wilhelm, Mathias; Van Damme, Petra; Menschaert, Gerben: PROTEOFORMER 2.0: Further Developments in the Ribosome Profiling-assisted Proteogenomic Hunt for New Proteoforms. Molecular & Cellular Proteomics 18 (8), 2019, S126-S140 mehr…

Tools utilizing Prosit

Oktoberfest

Oktoberfest is a python tool for collision energy calibration, rescoring search results and generating spectral libraries for proteomics research within the Prosit ecosystem. It offers an end to end pipeline that takes search results, predicts peptide properties using koina, plots summaries and quality control figures and performs FDR estimation with either mokapot or percolator.

Koina

Koina is a community driven endeavour to make AI models for life science research accessible. It is a service that tries to reduce the barrier of entry for using machine learning models in life science research by removing the need to set up and maintain the infrastructure required to run the models. The project is open-source and adheres to FAIR principles.