Periodic Reporting for period 2 - PROTrEIN (Computational Proteomics Training European Innovative Network)
Période du rapport: 2023-01-01 au 2024-12-31
Mass spectrometry-based proteomics is one of the most powerful technologies for the study of the cellular proteome, protein interactions, and post-translational modifications. Although several proteomics workflows are well-established methods, the field is fast evolving with new acquisition approaches and the generation of high-content data structures. This evolution has expanded the number of applications, and mass spectrometry-based proteomics has become one of the principal platforms in the multi-omics universe. However, many of these developments require the assistance of machine-learning approaches to predict peptide properties and optimize the acquisition of proteomics datasets. These new applications often require the development of new processing algorithms to query and interpret raw mass spectrometry data, and therefore reveal hidden proteoforms, interactions, and protein modifications. The implementation of the new experimental possibilities also requires the development of interactive and highly-visual tools for the integration and representation of proteomics data with other omics datasets. Addressing these different needs is crucial to use proteomics at its full potential and thus deliver major contributions to our understanding of biological processes as well as to support the health sector with valuable information. Within this context, we need a new generation of researchers that can delve into the complexity of the computational proteomics field. Indeed, institutions and funding bodies should carve out a viable place for bioinformaticians who focus on collaborations and reward them for their abilities to navigate the myriad demands of multidisciplinary projects. PROTrEIN nurtures this new generation of computational proteomics scientists by training them on the mastering of i) the mass spectrometry raw data nature, structure, and information content, ii) mass spectrometry acquisition methods, iii) algorithms, gamification, and machine learning tools, iv) data integration and visualization, v) science communication, vi) data management, and vii) research integrity and innovation (RRI).
To successfully achieve PROTrEIN´s goals, we designed a comprehensive research program around PROTrEIN´s three main Research objectives. The success of the project – where PROTrEIN ESRs made crucial contributions in achieving and overdelivering on the network´s goals, as well as important contributions to basic research, personalized medicine, biomarker discovery, protein structure analysis, and therapeutics development - is a testament to the dedication of all partners, the excellence of the training and supervision, and the fruitful collaborations forged.
In Research Objective 1, several AI-driven tools were developed to enhance computational proteomics. These tools included a deep learning model to predict peptide stability; the XMass model, that improves spectral library generation ; the iDeepLC model that significantly improves peptide retention time prediction for modified peptides; and Prosit-XL, which accurately predicts fragment ion intensities for cross-linked peptides. These advancements have accelerated proteomics workflows, enabling more reliable and precise analyses and providing researchers with enhanced tools to uncover deeper insights into biological processes and disease mechanisms.
In Research Objective 2, computational advancements were made to improve mass spectrometry data analysis and expand the detection of novel proteoforms. Software tools were developed to address long-standing challenges in processing raw MS spectra, providing powerful new workflows for detecting hidden proteoforms. Furthermore, new open-search strategies and integrated scoring pipelines were introduced to help identify and quantify phosphorylation sites; and updated cross-linking search engines enabled more confident detection of protein–protein and protein–nucleic acid interactions.
In Research Objective 3, computational tools were developed to enhance multi-omics visualization and biological interpretation.A Perseus module integrated proteomic and fMRI data, enabling analysis of protein behavior in relation to brain activity. A circadian rhythm analysis toolkit was introduced; and the NuXL search engine was integrated into Proteome Discoverer, improving protein-nucleic acid complex analysis.Multi-omics data visualization was implemented in RDConnect, aiding rare disease diagnosis in 216 individuals. Network-based algorithms were developed to link proteomes to phenotypes, and a web-based game was created to teach protein interactions and drug mechanisms. A VR experience, "Seen the Unseen," was developed to illustrate molecular interactions, with further gamified tools in progress.
All PROTrEIN fellows presented their work at international conferences and meetings, and several manuscripts have been published, with many more in the pipeline. The whole consortium was involved in communication and outreach activities, such as social media campaigns, Wikipedia editathons , and the PROTrEIN art installation.
Innovation and entrepreneurship skills were also an important focus. PROTrEIN’s training program practical and theoretical training to help ESRs develop business ideas, participate in a pitching competition, and receive expert feedback on translating research into viable business ventures.
In Research Objective 1, several AI-driven tools were developed to enhance computational proteomics. These tools included a deep learning model to predict peptide stability; the XMass model, that improves spectral library generation ; the iDeepLC model that significantly improves peptide retention time prediction for modified peptides; and Prosit-XL, which accurately predicts fragment ion intensities for cross-linked peptides. These advancements have accelerated proteomics workflows, enabling more reliable and precise analyses and providing researchers with enhanced tools to uncover deeper insights into biological processes and disease mechanisms.
In Research Objective 2, computational advancements were made to improve mass spectrometry data analysis and expand the detection of novel proteoforms. Software tools were developed to address long-standing challenges in processing raw MS spectra, providing powerful new workflows for detecting hidden proteoforms. Furthermore, new open-search strategies and integrated scoring pipelines were introduced to help identify and quantify phosphorylation sites; and updated cross-linking search engines enabled more confident detection of protein–protein and protein–nucleic acid interactions.
In Research Objective 3, computational tools were developed to enhance multi-omics visualization and biological interpretation.A Perseus module integrated proteomic and fMRI data, enabling analysis of protein behavior in relation to brain activity. A circadian rhythm analysis toolkit was introduced; and the NuXL search engine was integrated into Proteome Discoverer, improving protein-nucleic acid complex analysis.Multi-omics data visualization was implemented in RDConnect, aiding rare disease diagnosis in 216 individuals. Network-based algorithms were developed to link proteomes to phenotypes, and a web-based game was created to teach protein interactions and drug mechanisms. A VR experience, "Seen the Unseen," was developed to illustrate molecular interactions, with further gamified tools in progress.
All PROTrEIN fellows presented their work at international conferences and meetings, and several manuscripts have been published, with many more in the pipeline. The whole consortium was involved in communication and outreach activities, such as social media campaigns, Wikipedia editathons , and the PROTrEIN art installation.
Innovation and entrepreneurship skills were also an important focus. PROTrEIN’s training program practical and theoretical training to help ESRs develop business ideas, participate in a pitching competition, and receive expert feedback on translating research into viable business ventures.
The PROTrEIN network addressed the demand for skilled computational proteomics researchers in Europe by integrating machine learning, bioinformatics, and MS-based proteomics. It provided cross-sectoral doctoral training, combining academic coursework with non-academic placements to foster collaboration among universities, research institutes, and industry. By enhancing innovation in proteomics and emphasizing digital fluency and career preparedness, PROTrEIN helped set a new standard for doctoral training in Europe.
The network’s curriculum aligned with the EU’s innovation goals and the Future of Jobs Report 2025, targeting sills in AI, big data, and creativity. ESRs received specialized training in data science, bioinformatics, project management, data ethics, open science, and leadership, alongside entrepreneurship modules. Secondments at non-academic partners provided hands-on experience, while career sessions covered entrepreneurship, communication, and IP management. These efforts prepared graduates to apply data-driven skills in biomedicine, biotechnology, and pharma, strengthening Europe’s innovation workforce.
By advancing the frontiers of computational proteomics, PROTrEIN bolstered Europe’s leadership in precision medicine, biomarker discovery, and data-driven approaches to healthcare. Scientifically, the network’s broad collaboration enabled the development of novel AI tools that improved peptide and protein identification, increasing both accuracy and throughput of MS-based workflows. These innovations will potentially find applications, e.g. in clinical research, accelerating the creation of reliable diagnostics and more targeted therapies.
The network’s curriculum aligned with the EU’s innovation goals and the Future of Jobs Report 2025, targeting sills in AI, big data, and creativity. ESRs received specialized training in data science, bioinformatics, project management, data ethics, open science, and leadership, alongside entrepreneurship modules. Secondments at non-academic partners provided hands-on experience, while career sessions covered entrepreneurship, communication, and IP management. These efforts prepared graduates to apply data-driven skills in biomedicine, biotechnology, and pharma, strengthening Europe’s innovation workforce.
By advancing the frontiers of computational proteomics, PROTrEIN bolstered Europe’s leadership in precision medicine, biomarker discovery, and data-driven approaches to healthcare. Scientifically, the network’s broad collaboration enabled the development of novel AI tools that improved peptide and protein identification, increasing both accuracy and throughput of MS-based workflows. These innovations will potentially find applications, e.g. in clinical research, accelerating the creation of reliable diagnostics and more targeted therapies.