Embedding the complexity of disease into the drug discovery process
Many of the most powerful drugs act via multiple modes of action.
Aspirin, the fever and headache medication, reduces inflammation in your body via two different but complementary pathways. By engaging multiple biological targets, aspirin is the first line of treatment for a wide variety conditions, and is one of the most important medicines discovered. Importantly, aspirin is also one of the safest drugs; its many interactions across the body do not cause additional toxicity.
Since the synthesis of aspirin in 1899, modern drug discovery has moved towards designing therapeutics for specific molecular targets. With the advent of sequencing technology in the 1980s and advances in molecular biology, researchers have been able to pinpoint proteins involved in the pathology of certain diseases. Mutations in genes such as CFTR, HTT, and HBB manifest maladies such as Cystic fibrosis, Huntington’s disease, and Sickle Cell Anemia, respectively. In the past forty years, modern drug discovery has built a suite of computational and experimental tools that can find drugs against single, specific proteins.
While one-drug one-target ‘magic bullets’ are often life-saving therapeutics; focusing on just specific molecular targets ignores the fact that most diseases are complex — driven by an amalgam of distinct processes rather than by just a single protein.
Diseases are often caused by a variety of mechanisms: combinations of proteins being dysregulated, mutations across several genes, distinct cellular systems going haywire. While magic bullets have been incredibly successful for the past twenty years of drug discovery, we are now seeing the limitations of that approach. Issues such as lack of efficacy, drug resistance, and on-target toxicity of modern therapeutics can be attributed, in part, to a failure of embedding the complexity of disease into the drug discovery process.
While we need therapeutics that address the cacophony of disease mechanisms, our current arsenal for designing these types of drugs is woefully inadequate. Nowhere is this clearer than for drugs targeting a family of proteins called kinases.
Kinases and the complexity of their inhibitors
Kinase inhibitors are precision medicines, designed to target a single kinase protein implicated in diseases like cancer or autoimmunity. Coincidentally, these precise drugs operate by a multitude of other mechanisms as well. Kinase inhibitors can, for instance, inhibit both intended kinases (on-targets) and non-specific kinases (off-targets), engage mutated versions of a disease-causing kinase, or even change a kinase’s shape to disrupt protein-protein interactions.
Administration of kinase inhibitors results in strong therapeutic effects, such as deactivating cancer resistance mechanisms or suppressing an overactive immune system. But their mechanisms of action are not yet entirely clear, as kinase inhibitors may simultaneously cause toxicity as severe as cardiac arrhythmias and brain bleeding. This discordance between efficacy and safety significantly limits kinase drugs from treating a wide variety of diseases. At Harmonic Discovery, we see an enormous opportunity to create safer and more efficacious drugs; our mission is to unlock the full potential of kinase therapeutics.
Unlocking the full potential of kinase therapeutics
We are building a machine learning-first infrastructure that harmonizes all aspects of kinase drug discovery. Our models learn from the topology of atoms and bonds that make up a kinase drug, to the 3-dimensional ins and outs of kinase protein structure, all the way to the dynamic cellular environments of disease driven by kinases. Our integrated approach enables us to create molecules where we have control not only over the scope of kinases we target, but also the extent to which they are inhibited. By controlling the scope and the extent of kinase inhibition, our vision is to create a new generation of therapeutics.
In the near future we envision the discovery of medicines that have lower toxicity arising from heavily inhibiting a single protein. Or cancer drugs that can overcome several resistance mutations seen in a single kinase, or across multiple kinases. With the advent of massive computational power, as well as advances such as CRISPR gene editing, we believe that drug discovery spanning any combination of kinases across the kinome and, eventually, any combination of genes in the whole genome, is just over the horizon.
We are honored to have the support of several partners that embrace our vision for the future of medicine. These include institutional investors such as Innovation Endeavors that led our seed financing round, along with Fifty Years, Y Combinator, Boom Capital, and Caffeinated Capital, our angels, Ethan Perlstein and Harry Glorikian, and partners such as the Johnson and Johnson Innovation Labs, ProHealth and TechParks Arizona.
We are also tremendously grateful for the advice from our scientific advisory board members: Drs. Olivier Elemento, Omar Abel-Wahab, Tero Aittokallio, and Andreas Bender.
Finally, we are thrilled to announce that Dr. Joel Dudley will be joining Harmonic Discovery’s Board of Directors. Dr. Dudley joins Harmonic Discovery from Tempus Labs, where he was Chief Scientific Officer. Prior to Tempus, Dr. Dudley was the Director of the Institute for Next Generation Healthcare at the Mount Sinai School of Medicine.
The science behind Harmonic Discovery is derived from the brightest minds across the world in machine learning and drug discovery. We are launching our company with three oral presentations at the American Chemical Society Fall 2022 conference introducing key aspects of our platform.
We are still looking for incredible people that believe in our vision. If you are interested in joining the team, click here.
Rayees, Jason, and Marcel
