And this is right after design and low-volume synthesis, relatively cheap and early parts of the process.Ī Manifold Bio employee works in their wet lab. This screening process takes a lot of uncertainty out of the equation, since you know for sure that this protein actually does what you designed it to do, in a living mammalian system. That's 98 molecules you don't have to take through further testing. Equal quantities of 100 molecules go into the mouse, maybe 95 don't do anything, three attach to the cancer decently and two attach at a much higher rate. Each protein, however similar, gets a tag that is totally unique and identifiable after the fact when put through a proprietary DNA conversion process. That's where Manifold's barcodes come in. The problem here is that at this stage, you might be looking at a thousand different proteins, each of which may only vary from the others by a couple amino acids, the building blocks that give proteins their shape and function. We focus on a specific cancer target, something on the surface of cancer cells that really flags that this is a cancer cell, and figure out the drug designs that go very specifically to those cancer cells - and don't go anywhere else.' There's a lot of AI and machine learning- driven development, a lot of design on the computer. Rather than verifying lab tests, 'It's more for informing design. 'The way it works is we attach an extra bit of protein, our protein barcode, and that makes the protein trackable and we can track it wherever it goes in a system,' Kuznetsov said. Illustration of multiple 'barcodes' attached to proteins to be tested in vivo. But it also is the first - and often the last - time any of these drugs actually has to work in a real body, and the result is many washing out. Because this is such a logistically challenging and time-consuming part of drug testing, it is usually left until there are only a handful of molecules the company or lab is fairly confident will work. And the general equation there is: one drug, one mouse. In vivo means 'in life,' as opposed to in vitro (in glass) or in silico (simulated), and mice are the usual critters that get the dubious honor of testing out a new drug's safety and efficacy. You don't have to know what all of those do, however, to understand that in the end, no matter how much they accelerate fundamental research, in vivo testing remains a major bottleneck. The drug discovery space has advanced a lot in the last few years, first with fast and cheap gene transcription, then with CRISPR then with AI-powered proteomics.

Manifold Bio's molecular machinery could let a hundred molecules be tested simultaneously in a single living system, potentially upending the whole process.

In the creation of a new drug, there comes a point where you have to finally put the molecule in a real creature - one at a time - and see if it actually does what you think it does.

Manifold bio