FAU researchers aim to improve the diagnosis and treatment of both diseases
Epidemiological studies indicate that people with inflammatory bowel disease have a lightly increased risk of contracting Parkinson’s later in life. The reasons for this are still unclear. Dr. Pooja Gupta, group leader of “Bioinformatics for Multi-Omics” at Uniklinikum Erlangen will pursue this question in the coming five years. Her project “AI-PREDICT” uses artificial intelligence to investigate the potential biological connections between the two diseases. The Federal Ministry of Research, Technology and Space (BMFTR) is providing approximately two million euros in funding for the initiative.
The intestines are home to an enormous variety of bacteria, viruses and fungi – a similar number of them to cells in the human body. This complex ecosystem with its metabolic products and signaling molecules has a decisive impact on the brain. There are indications that people with chronic inflammatory bowel disease (IBD) contract Parkinson’s disease more frequently. “In our project we aim to identify molecular patterns that occur with both diseases,” says Gupta. “In this way, we hope, for instance, to be able to identify people with an increased risk for Parkinson’s at an earlier stage.”
Molecular interactions between gut and brain
The exact molecular interactions between IBD and Parkinson’s are still unclear. In order to close this gap, Gupta’s team will use artificial intelligence (AI) to investigate several biological levels at once, from the microbiome in the digestive tracts of the patients to genes, proteins and metabolic products that arise particularly frequently in both diseases. By jointly analyzing these levels, the team hopes to disclose inflammation signatures that affect both the gut and the brain. In this way, they hope to gain an insight into how IBD and Parkinson’s arise and understand why they often present as comorbidities.
“We will use innovative deep-learning techniques such as graphic neural networks to identify these fingerprints,” Gupta explains. “We can use them to analyze how various biological molecules interact.” This allows the AI to group the patients not only according to their symptoms but also according to specific molecular patterns of their disease. “We also hope that the signatures we discover will allow us to predict which people with IBD may have an increased risk of developing Parkinson’s years before the first symptoms appear,” Gupta continues. Findings from the project could make a contribution towards proactively delaying or preventing the onset of Parkinson’s disease.
The system will be trained using data from IBD and Parkinson’s patients. The patients are from Erlangen (Prof. Dr. Markus Neurath, Department of Medicine 1 and Prof. Dr. Jürgen Winkler, Department of Molecular Neurology) and from international cohorts. One major challenge is harmonizing the various sources in order to allow for analyses that can be compared and stand up to robust interpretation.
A blueprint for the medicine of the future
The impact of AI-PREDICT will go beyond IBD and Parkinson’s. By modeling the interactions between various organ systems, the project can provide a blueprint for investigating other comorbidities. In medicine, these are cases in which patients suffer from several connected diseases.
In order to ensure the success of the project, Gupta is counting on an interdisciplinary network of clinical experts and computer specialists. The project is based at the Department of Stem Cell Biology (head of department: Prof. Dr. Beate Winner).
Further information:
Dr. Pooja Gupta
Head of the working group “Bioinformatics for Multi-Omics”
Uniklinikum Erlangen
Phone: +49 9131/85-42840
Email: pooja.gupta@uk-erlangen.de