GGrantIndex
← Search

Generative AI-guided design of HIV-1 cocktail immunogen

$684,992ZIAFY2025AINIH

National Institute Of Allergy And Infectious Diseases

Investigators

Abstract

This initiative seeks to incorporate generative AI technologies into the development of HIV-1 vaccines with the objective of expediting the process. This marks a notable departure from conventional rational methodologies. The proposed approach utilizes advanced protein diffusion models, graph neural networks, and large language models to create and assess novel mini-protein immunogens derived from identified epitope hotspots. Initial efforts have effectively employed these technologies to tackle the structural and genetic variability of HIV-1 and other viral proteins by designing constructs that replicate essential neutralizing epitopes. A primary emphasis is placed on the creation and validation of de novo mini-proteins that faithfully reflect the natural structural features of the HIV-1 Env trimer. This entails generating an extensive library of 10,000 unique scaffolds informed by motif variations and utilizing AI models such as AlphaFold2 for in silico validation. Following this, experimental assays will focus on the most promising designs based on their binding affinity, stability, and structural alignment, employing advanced techniques like Biacore and CryoEM. Furthermore, the project investigates nanoparticle-based systems for immunogen delivery. By conjugating mini-protein immunogens to self-assembling nanoparticles, particularly mosaic nanoparticles, the strategy aims to provoke broad immune responses through the presentation of diverse epitope variations. These engineered immunogens and nanoparticles will be prepared for forthcoming pre-clinical trials to evaluate their immunogenicity in animal models, thereby laying the groundwork for potential clinical applications. In summary, this pioneering approach is dedicated to accelerating the development of effective HIV-1 vaccines by merging computational accuracy with experimental thoroughness, thereby offering promising pathways for future pandemic preparedness.

View original record on NIH RePORTER →