By Kenichi Hartman, Ph.D.
I've been seeing a lot of press lately about a new US/Israeli startup called Immunai, which recently raised a $20 million seed round led by Viola Ventures and TLV Partners. Now, $20M for a seed round is very impressive, and I wanted to know what they were all about.
Not surprisingly, what I found in the tech press and the company's website was long on hype and short on actual tech. One business publication declared that "Immunai has developed a vertically-integrated platform for multi-omic single-cell profiling that offers a broader view of the immune system in states of health, disease, and treatment to examine the body’s response to various treatments." This sentence is one of those journalistic sleights of hand that sounds technical but is vague to the point of meaninglessness to anyone with basic familiarity with cell and molecular biology. It also sounds like it could be Theranos-level world-conquering BS.
So I decided to look for patents. The company was founded in December 2018, so maybe they already have a patent application that's gone through its 18-month publication? No dice. A quick check in WIPO's Patentscope for an application from Immunai found nothing.
Another possibility, given that the founding team includes Prof. Ansuman Satpathy of Stanford and Prof. Dan Littman of NYU, is that Immunai's platform tech is covered by academic IP that's been licensed to them.
Bingo. Here's what I found:
PCT application PCT/US2018/066592 "Systems and methods for multiplexed measurements in single and ensemble cells" from Stanford, with Prof. Satpathy as an inventor. (https://patents.google.com/patent/WO2019152108A1)
This application summarizes the invention as follows:
A method of processing immune cells, comprising: (a) capturing an immune cell, wherein the immune cell comprises genomic deoxyribonucleic acid (gDNA) and messenger ribonucleic acid (mRNA) molecules; (b) contacting gDNA from the immune cell with a transposase to generate tagged gDNA fragments in a tagmentation reaction; and (c) generating complementary DNA (cDNA) molecules from the mRNA molecules, wherein the cDNA molecules comprise sequences that correspond to a V(D)J region of a genome of the immune cell.
Now this is something I can sink my teeth into. This sounds like actual tech, and you know what, it doesn't sound like BS either (having a Nature Medicine paper on this also helps https://www.nature.com/articles/s41591-018-0008-8). It also filters out the tech-press hype language that tends obfuscate more than enlighten. For one thing, "vertically-integrated platform for multi-omic single-cell profiling" sounds like Immunai might be working on a magic single cell next generation sequencing platform that will go head to head with Roche and Illumina. They are not. Rather, a chunk of that $20M will likely go to buying and running Illumina sequencers.
So here's the short but technologically more meaningful take: Immunai has licensed core tech from Stanford on how to build single-cell profiles of immune cells (e.g. B-cells and T-cells) that includes their genomic profile (from gDNA), transcriptomic profile (from cDNA), and source patient health information. They will build a large immune cell profile database from thousands of samples, then apply machine learning (lots and lots of clustering?) to the collected profiles to hopefully extract key features that are predictive of health outcomes, such as receptivity to a therapeutic agent.