AI Transforming Labor Challenges in Rare Disease Treatment
Despite advancements in modern biotechnology, thousands of rare diseases continue to lack effective treatments, largely due to a shortage of skilled professionals in the field. At the Web Summit Qatar, industry leaders from Insilico Medicine and GenEditBio highlighted the emerging role of artificial intelligence (AI) as a game-changer in addressing these labor challenges.
Insilico’s CEO Alex Aliper discussed the company’s vision to create “pharmaceutical superintelligence.” Recently, they unveiled the “MMAI Gym,” designed to enhance generalist large language models like ChatGPT and Gemini to operate with the same level of efficacy as specialized models. This innovative platform aims to develop a versatile AI capable of tackling multiple drug discovery tasks with superhuman precision.
The Insilico platform leverages biological, chemical, and clinical data to formulate hypotheses regarding disease targets and potential therapeutic candidates. By automating processes that traditionally required extensive manual efforts, the platform can efficiently navigate expansive design spaces, identify high-quality therapeutic options, and even repurpose existing medications, all while significantly reducing costs and timeframes. For instance, the AI recently identified potential avenues for repurposing existing drugs for treating ALS, a rare neurological condition.
However, the labor shortage extends beyond the drug discovery stage. Effective treatment often necessitates interventions at fundamental biological levels. GenEditBio is pushing the boundaries of CRISPR gene editing through its innovative strategies. This “second wave” focuses on achieving precise in vivo gene delivery, enabling straightforward one-time injections directly into affected tissues.
GenEditBio’s strategy hinges on a robust database of unique, non-viral polymer nanoparticles designed to efficiently deliver gene-editing tools. Utilizing its NanoGalaxy platform, the company applies AI to correlate chemical structures with distinct tissue targets, such as the eye or liver, predicting necessary adjustments to delivery vehicles that minimize immune responses. Their iterative testing process refines these predictions, ensuring precise, tissue-targeted deliveries vital for successful in vivo gene editing.
Zhu, a representative from GenEditBio, asserts that their approach not only diminishes costs but also paves the way for off-the-shelf therapies that enhance patient accessibility. Recently, the company secured FDA approval to initiate trials for CRISPR therapy targeting corneal dystrophy.
Addressing the persistent data challenges is crucial for optimizing AI systems within biotech. Aliper emphasized the need for broader datasets, noting that most current research originates from the Western world, leading to inherent biases. Insilico’s automated labs produce extensive biological data from disease samples, feeding this information into their AI platform for enhanced discovery capabilities.
Zhu highlighted that valuable data are inherently present in the human genome, shaped by centuries of evolution yet often difficult for researchers to interpret. Both Insilico and GenEditBio are working towards harnessing these insights through advanced AI modeling techniques.
Looking ahead, Aliper mentioned plans to develop digital twins of humans to conduct virtual clinical trials. This initiative aims to move beyond the current stagnation of approximately 50 FDA drug approvals annually, especially in light of the rising prevalence of chronic conditions due to global aging. He expressed optimism that, within the next decade or two, we could witness a significant increase in personalized treatment options.
