Marcelo Paez-Pereda

Digital innovation has proven to be a transformative force, revolutionizing processes from drug discovery to real-world studies. The application of deep-learning techniques has paved the way for precision medicine, enabling the analysis of multi-omics data and fostering the era of personalized medicine. From generating novel molecular structures to simulating drug-target interactions, digital tools have become indispensable in advancing medical insights and optimizing drug designs.

In the realm of rare diseases, however, the utilization of these ground-breaking technologies presents unique challenges. Rare diseases, are inherently complex and poorly understood, affecting over 300 million people worldwide. Unfortunately, a staggering 95% of rare diseases lack approved or effective treatments. The disparity in resources allocated to rare diseases, compared to more prevalent conditions, exacerbates the challenges, leading to scarcity or non-existence of real-world data (RWD) and less precise clinical endpoints.

One of the primary hurdles for leveraging digital to advance the rare disease field - is the scarcity and fragmentation of data. Unlike more common diseases, data related to rare diseases are not only limited but also scattered across various databases and sources. The low prevalence of rare diseases, coupled with delays in diagnosis and the absence of effective treatments, compounds the difficulty in developing a comprehensive understanding of key disease phenotypes and progression.

To address some of these challenges, our digital team has first focused on de-risking clinical trials early on, with a particular focus on translational sciences . The foremost challenge we are tackling is the unbiased estimation of optimal animal models that faithfully replicate human disease conditions. Animal models are indispensable for unraveling the genetic bases and molecular mechanisms of rare diseases. By leveraging advanced analytic techniques and public data “ we were able to create an automated platform to support translational teams in identifying optimal models with the highest translational potential.

A paradigm shift in our approach also involves viewing rare diseases not in isolation but as groups with shared clinical features. This shift enables us to collectively address the unmet needs of various rare diseases, optimizing resources and fostering a more collaborative environment and paving the way for basket trials. By considering diseases in groups rather than isolation, we aim to identify phenotypic similarities across different diseases, ultimately leading to a more profound understanding of disease pathophysiology.

We are exploring the concept of disease adjacencies, defined by shared phenotypes in corresponding animal models. Evaluating shared features among indications, allows for resource optimization by grouping and scoring indications with similar features together. This novel approach, backed by machine learning models, seeks not only to determine phenotypic similarities across different animal models of various indications, but also enables us to determine similarities across human rare diseases.

By addressing the challenges of data scarcity, lack of understanding, and the isolation of rare diseases, our approach aims to bring more collaboration and efficiency in the pursuit of simultaneously addressing multiple diseases with high unmet needs. Through the strategic use of digital technologies and close partnership with translational and clinical teams, we aim to not only advance rare disease research but also pave the way for a more interconnected and informed healthcare landscape.