The future has arrived. The laborious process of drug discovery, often fueled by genius and serendipity has now been captured on a platform powered by artificial intelligence (AI). Scipher Medicine’s Spectra platform leverages the human interactome and AI to analyze large patient molecular data sets to decipher the complexity of disease biology.
Slava Akmaev, PhD, chief technology officer and head of therapeutics at Scipher has been named one of the top AI leaders in drug discovery and advanced healthcare. Here he describes the Spectra platform and its network map of human biology and how it can be used to decipher the causal relationship between proteins expressed from the human genome and disease phenotypes. The result is a new type of genius, one less dependent on serendipity, and powered by the technology of the future.
Spectra is a breakthrough because it allows us to rapidly, within a number of months, identify and prioritize the entire human proteome as potential therapeutic interventions as it relates to a specific complex disease.
Scipher’s founder, Laszlo Barabasi at Northeastern University in Boston had spent more than a decade understanding pairwise protein-protein physical interactions and building the most comprehensive map or the human interactome of such interactions in the human cell. The protein-protein interactions are critical biochemical processes in the cell and often represent the underlying mechanisms of molecular dysregulation in a particular disease. Scipher licensed this technology from Northeastern University and built the Spectra platform using the human interactome as its foundation.
Spectra is built to address two critical questions in life sciences. One, can Spectra identify molecular biomarkers for patient stratification to a better treatment strategy. For example, Scipher recently launched a diagnostic test, PrismRA® that uses a set of Spectra discovered biomarkers predictive of inadequate response to anti-tumor necrosis factor (TNF) therapy in patients with rheumatoid arthritis.
Two, can Spectra discover novel therapeutic. An example of this is our application of Spectra to rank and identify novel targets in a different auto-immune disease, ulcerative colitis.
The company has partnered with the biotechnology company Galapagos to develop novel therapeutic targets identified by Spectra with the focus on ulcerative colitis. We look for specific biological networks implicated in the UC pathogenesis, rank the proteome according to the predicted efficacy of the target, and determine the most promising intervention points or genes to treat UC. Spectra ranks the proteome based on the likelihood of a possible intervention having an impact on the disease molecular dysregulation, i.e., reverting UC molecular pathophysiology back to the normal state. While Spectra looks for novel targets, it, of course, also presents the well-known targets on the priority list. Successful ranking of the known intervention points is in-silico validation of the technology.
In addition to identifying drug targets, Spectra can create a patient stratification strategy and identify biomarkers that may be representative of a particular patient molecular phenotype more likely to benefit from a specific therapeutic.
One of the concepts of science fiction is the ability to tailor treatments to specific clinical presentations and achieve cures. Right now, while the company does have cures for many diseases, Scipher also understands that many patients, oftentimes a majority of the patients do not respond to those interventions. This concept is even more critical for gene therapies where response/non-response can be such a discrete and unequivocal yes or no outcome. Spectra lies in the realm of science fiction because it may make it possible for us to always know what treatment or device or medical approach should be used for any given disease or adverse outcome.
Spectra, in combination with other automated drug development platforms, has the potential to disrupt drug discovery. Scipher Medicine reduced target discovery and validation times from years to months. With a Spectra enabled patient stratification strategy, drug human trials in a selected population can be focused on the patients with high likelihood of response, thus decreasing the N required to reach statistical significance. Consequently, it also dramatically reduces the time needed to complete clinical development. We hope to have new therapies on the market in under 10 years!
The point to emphasize is that now Scipher can develop therapies faster and at a much lower cost. This is a technological revolution.