Today’s post will focus on the vision for PLab. Let’s start with a basic statistical concept- long tail distribution. While this form of frequency distribution has been around since the 1940’s, it gained prominence in fields outside of insurance and finances, when it was popularized in a Wired article by Chris Anderson. You can read about this more here.
The long tail distribution has been applied to numerous business that focus on economies of scale. It is in this vein that it can be applied to PLab’s vision as well.
Take a look at this schematic here, the X axis represents diseases and the Y axis represents the number of people. Now, the bulk of pharmaceutical companies focus on diseases that are on the left of the dashed line, i.e. that affect huge population sizes. Intuitively, this makes sense, because more patients, means larger market sizes.
(Note: this is not to be critical of how companies make their decisions, but rather a reflection of status quo).
On the very same graph, the long tail region refers to diseases that have lower incidences and smaller patient populations associated with them, in other words, rare diseases.
Each of the rare diseases in this section do not offer access to large market sizes, but together, they allow one to reach a large population across multiple diseases.
How does this become feasible though? PLab uses a platform approach for parallel screening across model organisms and across multiple diseases.
Screening in model organisms is cheaper and allows for system-wide validation. For a single disease, the drug discovery process alone is reduced both in time and costs allowing us to achieve results at a third of the cost of traditional drug discovery. We’ve already seen this with the first disease in our pipeline, namely Niemann Pick Type C disease.
Further, with each additional disease, the cost of screening will go down so that with each added disease, the process of drug discovery becomes more efficient as well as economic.
So, what kind of diseases can PLab make a difference to?
There are about 7000 rare diseases, of which roughly 50% are heritable, i.e. relating to genes. Not all of these are monogenic and evolutionarily conserved, but a significant number of them have related pathologies. We have conservatively identified about 250 single gene diseases that we can access with our model organism platform. Each of these diseases in turn shares some overlap with the more common multigenic but heritable diseases, thus allowing us to indirectly query over a 1000 diseases. (E.g. Niemann Pick C disease shares some overlap and can be related to Alzheimer’s disease as well as hypercholesterolemia).
Thus, by targeting several rare diseases at once, not only can we find precise therapeutic solutions for diverse patient groups, we can also potentially find drugs that have therapeutic applications for diseases that affect large swatches of populations, i.e. a cure for the common from the rare.
Is this really feasible?
We believe it is! For NPC disease, we went from screening to lead molecule in under a year with version 1 of our screening platform. We’re in the process of designing and building a second version, PLab Platform V.2.0, that will have the following capabilities:
1) Sensor-based semi-automated husbandry and maintenance
2) Semi-automated phenotyping capabilities for our custom mutant models, incorporating image based data analytics
3) Database of mutational phenotypes and drug actions
4) In silico screening based predictions for target engagement with our drug libraries.
Future versions of the PLab platform will focus on additional layers of automation and scaling to help us achieve our goal of onboarding multiple diseases at once.