This story appeared in the department’s 2025 annual report
Research on antimicrobial peptides has come a long way in the past 45 years. And the Antimicrobial Peptide Database created by the laboratory of Guangshun Wang, PhD, a professor in UNMC’s Department of Pathology, Microbiology and Immunology, has been embraced around the globe by scientists, who have cited it thousands of times in their research.
The professor’s research focuses on innate immune antimicrobial peptides and their potential application as novel antimicrobials to combat drug-resistant pathogens. To promote research and education, his lab has been updating and expanding the Antimicrobial Peptide Database for two decades. This unique and comprehensive tool laid the foundation for his lab to decipher design principles of antimicrobial peptides and develop the database filtering technology for designing potent, selective and stable peptides with demonstrated efficacy in animal models.
Dr. Wang said that early in the database development, he wanted to understand natural peptides. “I thought if we understand that, even partially, we might be able to design better peptides. That’s why we limited our scope and initially restricted inclusion of synthetic peptides into our database.” The number of peptides discovered has increased through the succeeding versions of the APD, but it has been gradual, he said, because discovery of natural peptides is not easy.
After the international symposium in 2023 marking the 20th anniversary of the database, Dr. Wang and Dr. Joseph Khoury, department chairman, turned their eye to the future. “We started to think maybe we should expand the scope of data we can collect in this database” and make it even more useful. Therefore, this year brought a new version of the database, APD6. “After database reconfiguration, we can now add a representative set of synthetic peptides without losing the search power for natural antimicrobial peptides. The AI-predicted peptides will be included if they have actually been made in the lab and confirmed to have antimicrobial activity.”
The success of AlphaFold in protein structure prediction has stimulated the interest in predicting novel antibiotics in the same manner. While much progress has been made, Dr. Wang said there is a long way to go before AI-predicted peptides reach the market, calling the work in its early stages. “One of the major contributions of the APD6 is that we outlined a blueprint for advanced AI predictions. The database collects information for the entire information pipeline, from peptide discovery through all the procedures—at the end you will do in vivo assay, in vivo toxicity, then you will do efficacy and production, then you have clinical trials. Step by step, that’s the pipeline. And information for all these steps can be searched in the database.”
Further, Dr. Wang identified: “The data types can determine AI prediction quality. We don’t collect everything into the database. There’s a saying, garbage in, garbage out, or GIGO in computing science. We are doing it relatively conservatively by establishing a set of criteria for data registration.” The APD6 also proposes data filtering to remove noises, thereby improving AI prediction quality. Dr. Wang plans to report such filtering procedures in future database versions with the establishment of the AI prediction pipeline in collaboration with colleagues.
What’s new:
- Version 6 of the database contains more peptides, 5,680 as of September 2025, including 3,351 natural AMPs, 1,733 synthetic AMPs, and 329 AI-predicted AMPs.
- APD6 provides the first systematic classification of synthetic and predicted peptides.
- Also included in version 6 is an expanded wheel of peptide functions, which records 32 peptide functions or activities, such as antibacterial, anti-MRSA, anti-TB, antiviral, anti-HIV, anti-fungal, antiparasitic, anti-cancer, and anti-diabetic. The wheel of functions may open new opportunities for developing other types of peptide therapeutics in those areas.
- And an information pipeline (AMPIP) was created to facilitate future development of peptides in silico (by computer), thereby minimizing the use of research resources, including human labor and experimental animals.