Files and links (1)
PDFOpen Access (License Unspecified), Open Access
Thesis
MicroTraitLLM: a microbial-focused large language model for researchers
Master of Science (M.S.), Drexel University
May 2025
DOI:
https://doi.org/10.17918/00011083
Abstract
Large Language Models (LLMs) have become a focal point of biological and bioinformatics research in recent years. The vast number of applications, in combination with rapidly improving models, make LLMs an enticing point of investigation. However, LLMs often hallucinate if they are not provided with domain-specific information, which can misinform users. Additionally, with the vast number of applications being served in the eukaryotic domain, there has been little investigative research into LLM applications for prokaryotes, which has seen a surge in investigative research thanks to new high-throughput techniques. Furthermore, the current LLM tools for microbe-specific tasks provide insufficient citations, have a restrictive knowledge cutoff, or are too generalized for microbial researchers. Here we present MicroTraitLLM, a retrieval-augmented generation (RAG) system LLM which utilizes zero-shot and single-shot prompting to give specific, citation-based answers for researchers. Its connection to the PubMed Central live-updating, Open Access article database allows for the LLM to remain up-to-date on breakthroughs in the prokaryotic field. MicroTraitLLM can utilize various LLMs for its procedural answer generation, allowing for the user to customize their experience with the LLM they prefer the most. The tool is also instructed on citations, ensuring proper citations for a variety of formats. We show that MicroTraitLLM can respond in a similar time frame to many popular commercial LLMs, maintain relevant literature search, and provides informative answers to microbial experts.
Metrics
169 File views/ downloads
21 Record Views
Details
- Title
- MicroTraitLLM
- Creators
- Glen Rogers
- Contributors
- Gail L. Rosen (Advisor)
- Awarding Institution
- Drexel University
- Degree Awarded
- Master of Science (M.S.)
- Publisher
- Drexel University; Philadelphia, Pennsylvania
- Number of pages
- 14 pages
- Resource Type
- Thesis
- Language
- English
- Academic Unit
- School of Biomedical Engineering, Science, and Health Systems (1997-2026); Drexel University
- Other Identifier
- 991022058832804721