Progress and trends on machine learning in proteomics during 1997-2024: a bibliometric analysis.

Area of Science:

• Proteomics and Bioinformatics
• Computational Biology
• Data Science in Life Sciences

Background:

• Growing interest in machine learning (ML) applications within proteomics.
• Lack of a systematic mapping of the ML-driven proteomics research domain.
• Need to understand the knowledge structure, development trajectory, and emerging trends in this field.

Purpose of the Study:

• To conduct the first large-scale bibliometric analysis exclusively on machine learning-driven proteomics.
• To elucidate the knowledge structure, development trajectory, and emerging research trends.
• To identify key research foci, influential contributors, and collaboration patterns.

Main Methods:

• Bibliometric analysis of 5,156 publications (1997-2024) from Web of Science Core Collection.
• Utilized CiteSpace, VOSviewer, Scimago Graphica, and R bibliometrix package for data extraction and visualization.
• Analyzed keyword co-occurrence, citation networks, leading journals, authors, and institutional collaborations.

Main Results:

• Exponential publication growth since 2010, with a significant surge between 2019-2020.
• United States is the most productive country; Chinese Academy of Sciences leads institutions.
• Deep learning, particularly AlphaFold2, drives high citations; key themes include protein-protein interaction prediction and multi-omics analysis.
• Strong interdisciplinary convergence across computer science, molecular biology, and clinical research.

Conclusions:

• Presents the first comprehensive bibliometric overview of machine learning in proteomics.
• Highlights key themes: deep learning, pretrained models, and multi-omics integration.
• Recommends focus on interpretable models, enhanced collaboration, and standardized data use for advancing precision medicine.