The pharmaceutical industry still faces unrelenting challenges across the entirety of the drug development landscape. According to Piper Sandler’s recent report on “Artificial Intelligence and Machine Learning in Biopharma: From Hype, to Hope, to Impact,” biopharma is at a crossroads when it comes to drug development, underscored by huge expense, 90% failure rates, difficulties inherent to drug target identification and validation, a lack of robust biomarkers for patient stratification, and challenges in clinical trial design and execution. Decades from now, when we look back at this strange period in history, we think we will see that this was a critical inflection point for healthcare.
Artificial Intelligence (AI), first described in 1955 as the science and engineering of making intelligent computer programs, is weaving itself into all industries and quickly emerging as a significant disruptor in healthcare. Just follow the money: International Data Corporation (IDC) forecasts global spending worldwide for AI will jump from $85.3 billion in 2021 to more than $204 billion in 2025 with a 24% CAGR. Yet, irrespective of industry, one of the big challenges is integrating and deciphering large and complex datasets and applying that knowledge to practical situations to find solutions to real-world needs. Catalyzed by a perfect storm of the pandemic, technological advancements, and investment, the biomedical community seems to be on the precipice of unlocking the power of AI, with recent developments ranging from the creation of more detailed modeling of protein structures than ever before possible, huge progress in precision medicine, and the expansion of the use of Real-World Evidence in drug discovery and development.
Codifying expert-enabled AI into the drug hunting process
At HotSpot Therapeutics, we are passionate about finding meaningful therapeutic solutions for cancer patients and those battling auto-immune diseases, and we recognized the potential impact of integrating AI as a critical tool in various steps of the discovery and development paradigm. Simply put, we set out to create a platform that would take advantage of large datasets and use the analytical power of AI and Machine Learning as a core competency to make the discovery process more systematic, efficient, and potentially less costly. Creating this new toolkit will empower an acceleration of development by targeting the right patients and extracting the right patterns of information to inform signal seeking and future clinical design.
Over the last 36 months, HotSpot has diligently built and advanced our Smart Allostery™ platform using AI-enabled technologies to systematically identify, characterize, and ultimately drug so called “natural hotspots,” the privileged pockets on proteins that act as endogenous on/off switches. Staggeringly, only 2% of human proteins interact with currently approved drugs, but our platform allows us to attack undrugged and poorly druggable targets with allosteric drugs – an effort that is enabled through purposeful integration of AI applications. Briefly, we are concentrating HotSpot’s application of AI in three areas:
- Interrogation of large structure-function datasets to aid in identifying natural hotspots and characterizing their physical and molecular properties.
- Use of machine learning to interrogate the industry’s most extensive and diverse chemical library tailored to natural hotspots, which we established using internal DNA-encoded chemical library capabilities and an eye towards optimal drug-like properties.
- Exploring how to unlock the learnings of real-world clinical data to better enable precision medicine.
A strategic partnership to bring AI to precision medicine for difficult-to-drug therapeutic targets
As HotSpot prepares for the transition to a clinical-stage company, we believe we can accelerate the elucidation of drug development insights through deep interrogation of clinico-genomic data. Our recently announced strategic collaboration with Caris® Life Sciences is a linchpin in that strategy. Caris’s remarkable platform allows physicians and clinical researchers to assess the DNA and transcribed RNA products of 22,000 genes via next generation whole-exome sequencing and whole transcriptome sequencing. Our partnership is a unique opportunity to marry their self-described “where molecular science meets artificial intelligence” with novel allosteric medicines against difficult-to-drug targets.
This first-of-its-kind collaboration will reveal key clinical insights and enable more robust development decisions through access to robust real-world clinico-genomic data and real-time genomic analyses from our clinical trials. Specifically, we are focused on the following near-term capabilities:
- Interrogating a massive dataset with over 275,000 (and growing) cancer patient samples: HotSpot will have the ability to partner with Caris in asking critical translational questions of one of the world’s largest cancer patient datasets to find associations that can de-risk R&D decisions before human studies. The data will help drive decisions in our early development program, from hypothesis generation to hypothesis testing.
- Designing and executing more reliable and patient-centric clinical trials: As part of our Phase 1 studies, Caris will analyze liquid and tumor biopsies to generate comprehensive datasets of genomic and transcriptional data. While these data will be enormously important and the ability to inform decisions as we advance through clinical development, there is an inherent limitation in sample size. The ability to cross-compare these data back to Caris’ continuously growing patient database will give us incredible power to amplify early clinical signals to enable a much broader understanding of outcomes for precision medicine.
To conclude, it’s not always easy to see the historical importance of our actions in real time. Even the significant breakthroughs and profound moments of clarity we experience along the way seem to grow more refined with the perspective of hindsight. That is further confounded by the enormity of information that needs to be processed and thoughtfully actioned, a problem that AI applications can help us solve only when integrated into the bigger picture.
At HotSpot, we believe that the discoveries we are making right now are empowered by unprecedented technology, powerful data, and brilliant partners, and it is incumbent upon us as drug hunters to link our unlocking of protein biology to the development of safe and effective allosteric medicines for patients. Now is the time to seize that opportunity, and we are thrilled to have our new partners at Caris with us on that journey!