Syntensor is developing new machine learning methods coupled with first-principles thinking in biology to help drug developers, clinicians and physicians answer the fundamental question: Will it work?
Systems simulators have accelerated progress in every modernized field of engineering… except biology
Dynamical systems simulation tools have driven rapid innovation and growth in the aerospace industry and computing powered by advances in complex circuit design (see Moore's Law). Without equivalent tools, most drugs in clinical trials don’t make it to patients, while those that do may be ineffective or, at worst, harmful.
We’re on a mission to transform how medicines are developed and prescribed using advanced ML coupled with first-principles thinking in biology
The complexity and scale of the human biological system means ML methods are generally applied to only one part of it at a time: single cell or molecule RNA-Seq, protein-protein interaction networks or metabolic flux modeling, for example. However, results have shown that inferring the efficacy (or systemic toxicity) of a drug in a human from narrow biological sub-domains doesn’t work for drug development.
- 90% failure rate for drugs in clinical trials
- $1.3 billion to get a new drug to market
- 30-60% prescribed medicines are ineffective
- 4th leading cause of death in the US is adverse drug reactions
Instead, we need to accurately model the dynamics of the system, in order to predict the properties of the components within that system when they are perturbed by disease or drugs. The status quo is not sustainable. The total indirect cost of patient harm is estimated as $1 to 2 trillion globally per year.
We’re solving the 'complexity problem' in biology
Advances in machine learning methods for dynamical systems modeling and high performance computing mean it is now possible to model biological processes dynamically at an appropriate, systemic scale. Syntensor is the leader in the field and the first to apply these new fundamental methods commercially at scale.
Building a platform for every drug and all patients
The methods we’re using mean our platform is generalizable across diseases, therapeutic modalities and patient genotype/phenotype. Scientists and doctors will be able to simulate the effects of a drug on a patient to evaluate whether it will work and whether there are ‘red flags’ indicating off-target toxicity that could result in harm.
We’re a multi-disciplinary team of ML Researchers, Drug Discoverers, Software Engineers and Product Designers building a platform to predict drug efficacy and toxicity: a transformative tool for drug development and precision prescriptions.