AI-powered modeling approaches to support the development of new therapies for autoimmune diseases

Artificial Intelligence (AI) can support decision-making during drug development to select the right target, drug, dosing regimen and patient. AI and machine learning (ML) are useful to model disease heterogeneity, identify therapeutic targets within dysregulated molecular pathways, design and optimize drug-candidates, and evaluate clinical efficacy in silico. By creating predictive models on both the patient specificities and drug candidate properties, AI fosters the emergence of Computational Precision Medicine to better tailor therapies to the characteristics of individual patients in terms of their physiology, the pathophysiology of their disease and their susceptibilities to genetic and environmental risks.
This webinar will illustrate how, from the perspective of the pharmaceutical industry, various computational modeling strategies are being used to support the development of new treatments for primary Sjögren Syndrome (pSS) and Systemic Lupus Erythematosus (SLE), two autoimmune diseases with significant unmet medical needs. Multiomics profiling data of whole blood samples from hundreds of pSS patients and matched controls from the PRECISESADs IMI cohort were integrated to stratify patients by hierarchical and k-means clustering. A parallel modeling of pSS based on Artificial Neural Networks (ANN) data mining was undertaken by network computational analyses of transcriptomics data in blood and in salivary glands to identify therapeutic targets. In collaboration with ROSA, a quantitative system pharmacology (QSP) model of SLE was successfully developed to predict in silico the efficacy of the pan-neutralizing anti-interferon alpha S95021 monoclonal antibody.
Collectively, these various predictive models emerge as very powerful tools to inform drug development and support precision medicine strategies. They also provide supportive data to document drug efficacy and increase significantly the probability of success in future confirmatory real-world clinical studies.