The Evolving Landscape of Amyloid Research

Area of Science:

• Biomedical and Life Sciences
• Computational Biology
• Bibliometrics

Background:

• Exponential growth in biomedical literature challenges tracking discoveries.
• Artificial intelligence (AI) and natural language models (NLM) enable large-scale analysis of scientific publications.
• Understanding the evolution of specific research fields like amyloid biology is complex.

Purpose of the Study:

• To present a two-dimensional mapping of the global amyloid research landscape.
• To visualize key trends, historical progression, and dominant subfields in amyloid research.
• To identify latent research topics, including chemicals and small molecules, over time.

Main Methods:

• Utilized PubMedBERT, a transformer-based large language model.
• Employed t-SNE and Latent Dirichlet Allocation (LDA) for data analysis.
• Analyzed over 140,000 abstracts from the PubMed database.

Main Results:

• Provided a comprehensive visualization of the amyloid research field.
• Captured historical progression and the emergence of dominant subfields.
• Identified contributing authors, their countries, and latent research topics, including chemicals and small molecules.

Conclusions:

• AI-driven text analysis integrated with bibliometric data offers novel perspectives on amyloid research evolution.
• This study facilitates deeper interdisciplinary understanding of amyloid biology.
• Machine learning-driven literature mapping can identify knowledge gaps and guide future research.