Patient Science

Creating Medicines Specific for Each Genetic Disease

Patients Drive our Science

Many genetic diseases involve the loss of function of one or both copies of a gene leading to less functional protein being made within cells. This loss of function is what causes the imbalance in genetic disease. At Sharp we create small molecule drugs, the kind that can be taken as a pill. Our compounds restore the function of mutated proteins to correct the fundamental imbalance in each disease. With drugs that can treat a genetic disease at its core cause, diagnosis will be the first step on the path to better health and wellness.

We developed the CoreX™ biology, AlloChem™ chemistry and Disco™/Mine™ computational technologies to discover drugs that selectively activate the protein that’s been altered by gene mutation.

CoreX™ - Correcting Genetic Defects

CoreX is a suite of biology technologies that allows the creation of sensitive assays mimicking the disease state, enabling discovery of compounds that correct the mutational defect.

There are several mechanisms by which a small molecule can correct a genetic defect, as shown in the table below. All of these mechanisms have been demonstrated for small molecules.

Screen Shot 2021-05-05 at 11.26.30 AM.png

CoreX allows creation of assays to probe each of these mechanisms, and even to we take more than one approach to a given disease.

The CoreX suite includes biosensor technology licensed from the founders’ labs at Carnegie Mellon University along with proprietary assay methods developed at Sharp.

A key aspect of CoreX is that the assays have the sensitivity and signal-to-noise that makes large scale high-throughput compound screening possible.

AlloChem™ - Screening Libraries Enriched with Drug-Like Correctors

The AlloChem compound libraries are created using computational methods to select compounds most likely to bind proteins allosterically, thus stabilizing and activating them. Based on extensive supercomputing-scale thermodynamic simulations and analysis of known allosteric binding compounds, the AlloChem libraries are enriched with compounds that can functionally activate proteins while still complying with the standard rules for drug-like small molecules.

Screen Shot 2021-05-05 at 11.41.41 AM.png

These are not “virtual libraries” but are actual compounds that Sharp tests in real lab experiments in its CoreX assays to identify small molecules that correct genetic defects in the cellular context. While the compounds were selected computationally, the combination of the CoreX assays and the AlloChem libraries is an experimental discovery approach.

Disco™/Mine™- Big Data Analytics for Drug Discovery

Coming from Carnegie Mellon, of course we utilize best-in-class analytics!

Disco (short for “Discovery”) is a custom-built, cloud distributed analytics platform that captures all data for our AlloChem compounds in all CoreX assays. The platform allows custom CoreX assay data analysis enabling unique noise reduction and data integration for all CoreX assays.

Screen Shot 2021-05-05 at 12.08.41 PM.png

Disco gains knowledge with each new CoreX Assay run and each AlloChem compound tested. Disco employs the leading open system data analysis platform (Python and its associated libraries) enabling the implemention of Mine, an AI/Machine Learning platform to gain deeper insights into our experimental data. The purpose is not computation in place of experimentation, but rather experimentation augmented by computation.

NNFigure.jpg

Mine is implemented using Google’s TensorFlow and the PyTorch frameworks giving us the most robust platforms for bring AI/ML methods to bear on discovery data analysis. These tools allow us to discover drugs where traditional methods fail.

This triad of biology (CoreX), chemistry (AlloChem) and computational technologies (Disco/Mine) gives Sharp unique capabilities, significantly increasing the chances of delivering successful drug candidates for the treatment of genetic disorders.

To see how we’re applying the technology, visit our Priorities page.