AI model developed to bridge gap between animal models and human disease understanding

AI Model Developed to Bridge Gap Between Animal Models and Human Disease Understanding

Leipzig, Germany – November 2023 – In a significant stride towards refining the understanding and treatment of diseases, a new study published in the journal eBioMedicine details the creation of an advanced AI model. This endeavour, led by scientists from Leipzig University’s Institute for Medical Informatics, Statistics and Epidemiology (IMISE) and the Centre for Scalable Data Analytics and AI (ScaDS.AI), along with collaborators from the Department of Respiratory Medicine and Critical Care Medicine at Charité, aims to bridge the translational gap between animal models and human patients. The focus on COVID-19 showcases the potential impact of this innovative approach on clinical research and therapeutic development.

The scientific team harnessed single-cell RNA sequencing, a high-resolution method, to meticulously compare blood data from humans and various hamster species infected with COVID-19. The AI was trained using this molecular-level data to decode the differences and similarities in disease responses between the animal models and human patients.

Dr. Holger Kirsten, a scientist at IMISE and a corresponding author of the study, explains the breakthrough: “We have shown that the translational gap between animal models and human patients can be narrowed by integrating robust deep learning models in combination with biologically informed analyses. The AI systematically learns the molecular differences between the animal and the human, and can then translate the molecular patterns of the sick animal into corresponding patterns in humans, so to speak humanising the data from the animal model.”

The study identified specific insights into immune system responses in moderate cases of COVID-19. Dr. Geraldine Nouailles, Scientific Group Leader at the Department of Respiratory Medicine and Critical Care Medicine at Charité and another corresponding author, noted, “We were able to show that the activation of the immune system in moderate COVID-19 cases is very similar in Syrian hamsters and humans, particularly when considering monocytes.” Monocytes, as precursors to macrophages, play a critical role in the immune response.

Moreover, the research indicated that for investigating severe cases of COVID-19, Roborovski hamsters are a more suitable model due to the similar behaviour of neutrophils, the swift-reacting immune cells, in both Roborovski hamsters and humans. This alignment was significantly evident and is consistent with observations from human clinical data during the pandemic.

The implications of these findings are far-reaching. According to Dr. Kirsten, “Such comparisons of single-cell RNA sequencing data are well suited to reveal similarities and differences at the molecular and cellular level in animals and humans that go far beyond COVID-19 research.” This novel approach promises to enhance the identification of suitable animal models for human diseases, thereby improving the efficacy of preclinical studies and the precision of clinical trials.

Looking ahead, the Leipzig team intends to expand the application of this methodology to other diseases and animal models. A notable future direction includes investigating the effectiveness and safety of CAR T cell therapy, a promising treatment avenue for certain cancers.

The study, titled “Neural network-assisted humanisation of COVID-19 hamster transcriptomic data reveals matching severity states in human disease,” is accessible in the journal eBioMedicine. The findings mark a pivotal step in the evolving landscape of medical research, potentially transforming how translational studies are conducted.

More Information:
Vincent D. Friedrich et al, Neural network-assisted humanisation of COVID-19 hamster transcriptomic data reveals matching severity states in human disease, eBioMedicine (2024). DOI: 10.1016/j.ebiom.2024.105312

Source: Noah Wire Services