The project “Developing Novel Reliable Low Cost HIV Prognostics” is a German-Australian research collaboration between Prof. Dr. Dominik Heider and the Burnet Institute in Melbourne (Prof. Dr. Paul Gorry, Australia), funded by the DAAD. The aim of the project is to develop new, cost-effective and reliable prognostic methods for HIV patients using artificial intelligence (AI) and machine learning.

Human immunodeficiency virus (HIV) infects host cells via the CD4 receptor as well as a coreceptor (CCR5 or CXCR4). Since there is currently only one clinically relevant entry inhibitor (Maraviroc) for CCR5-tropic viruses, the determination of coreceptor tropism is essential for optimal therapy. However, currently available coreceptor prediction algorithms are primarily optimized for subtype B of HIV, even though it accounts for only 11% of global infections. The majority of HIV infections arise from subtypes C and A (together almost 75% of all cases), for which there are currently no reliable predictive models.

The project aims to develop a more reliable method for determining HIV coreceptor tropism for non-B subtypes using machine learning techniques and bioinformatics algorithms. Different analysis approaches are combined to improve prediction accuracy. The project is carried out as part of a German-Australian research cooperation and promotes scientific exchange between the two countries. The results are expected to contribute to the improvement of personalized HIV therapy and can be translated into clinical applications. The research project combines bioinformatics, machine learning and high-performance computing to develop a much-needed low-cost and reliable diagnostic system for HIV. The new method could revolutionize HIV therapy, especially for patients in countries with high proportions of non-B HIV subtypes.

Publications: doi: 10.2174/1570162x14666160321120232; doi: 10.1186/s13040-016-0089-1 

Contact: Univ.-Prof. Dr. Dominik Heider