Monday, November 24, 2014

Does anyone know what Cambrian Genomics' Austen Heinz is talking about?

I've covered Cambrian Genomics before. This is the Bay Area startup that claims to be "laser printing DNA." (Long story short: it's not actually "printing." Cambrian appears to be using lasers to sort DNA on microbeads.)

Cambrian's CEO, Austen Heinz, seems to have gotten himself in a little bit of trouble last week by claiming that his company's technology would enable (through making synthetic biology cheaper) sensitive body parts* to smell better. (Mr. Heinz has apparently dramatically recasted the original intent of another startup founder's idea.)

I'm really not interested in discussing that, since I don't think there's much value to be added from me/us to the discussion. However, I am very interested in figuring out if Austen Heinz's claims are true. Here's one from VentureBeat that I'd like to understand better: 
Cambrian’s technology is already being used for far less sensitive, and perhaps more useful, use cases. The company has been doing work printing DNA for the huge pharma company Glaxo Smith Kline. It’s also in talks to formalize a similar business relationship with Roche. 
Big pharma companies are asking Cambrian to print various types of DNA that can be used in the drug discovery and testing process. 
“We’re helping them make drugs,” Heinz said. For example, Cambrian’s DNA can be used for producing small molecules or for making new screens to find small molecules, Heinz said. “DNA can be used for every part of the process,” Heinz said. 
Heinz says that his company also intends to print DNA for customers in the industrial chemical and agricultural industries. He says producing seeds used by consumers is in itself a million-dollar industry.
So, first, I don't know what Mr. Heinz means by "Cambrian's DNA can be used for producing small molecules." Do they mean biologicals? (i.e. Cambrian is cloning genes that will make monoclonal antibodies?) I presume that "making new screens" is about using their technology to make protein for in vitro assays. 

I sure as hell don't know what Mr. Heinz is going to do with DNA for industrial chemical industries -- maybe this is a reference to biocatalysis? 

I also want to point out this interview with Planet Tech, where his predictions of what synthetic biology can do begins to wander into science fiction: 
...You recently described your vision of the future as "Anyone that has a mobile phone and bitcoin can create creatures." What exactly do you mean by this and how do you imagine it coming about?  
I mean that anyone with a phone can use genome design software for instance Benchling which runs on a web browser and order genes and DNA to make creatures that are useful to them.

Will people only be able to create new single-cell life or do you imaging the invention of entirely new larger organisms?

Yes i think new multicellular life forms built from a text file are possible but we need more progress in construction of large artificial chromosomes and the ability to print and sequence methylated dna at scale.

How do you plan to stop people from using your technology to create very dangerous microorganisms?

Virtualization. Instead of mailing out DNA we will send the DNA to a virtualization center like Transcriptic, Synthego, or Emerald Cloud Lab. From there they can put thousands of different DNA strands into thousands of cells then make thousands of video files of what those cells are doing and then do image process and machine learning on those videos and send that data back to the user to do the next design.

Not until the final organism is made will it be evaluated for release. This definitely lowers the bar for us for processing orders because as long as the screening is heavily locked down there is little risk of release of malicious code.
...Where do you see the future of the company in 5 and 10 years?

In the next 5 years we want to be the largest manufacturer of DNA in the world. In 10 years we hope to be closer to our longterm mission of replacing all natural organisms on the planet with better synthetic ones. For instance having made the DNA for say 10% of all plant on the planet surface sounds like a reasonable goal.
"Thousands of video files?" Does this even remotely make sense? Has there been some sort of advance in computer processing of microscopy files that I haven't heard about? (entirely possible) I know that he's predicting the future, speculating, "visioncasting", whatever. Count me highly, highly skeptical.

What I find really weird about the tech/venture journalism scene is how no one seems to be asking any scientists subject matter experts anyone if any of the stuff Austen Heinz says can happen is true or untrue:
  • Is it true that GSK has purchased DNA from Cambrian? 
  • Is it true that Cambrian is doing deals with Roche and ThermoFisher? (Somehow, I doubt it -- whose words are we relying on here? Mr. Heinz's, so far as I can tell.) 
  • He's obviously speculating about the future of synthetic biology -- how far off are his predictions from the median prediction of recognized experts in the field? 
Readers, what do you think?

UPDATE: Also, here are Cambrian Genomics' patent applications. 

*I'm not entirely a prude, just trying to write around corporate firewall software. 


  1. The science fiction part is still science fiction since life is not just about the parts, it is really about bootstrapping (the chicken or the egg problem). The virtualization bit is just a metaphor for something like Amazon Turk (hopefully with machines). However, the technology he is trying to develop is probably real. From the interview, he described it as:

    "DNA laser printing is laser catapulting of clonal DNA sequencing beads i.e metal beads, each metal bead covered in hundreds of thousands of identical copies of DNA. Essentially we create an explosion on the surface of the glass that projects our bead and associated dna into a tube where we can amplify and purify it for downstream use."

    I believe he is trying to build (may have succeeded) a micro-sized dna library using laser as micro-manipulator. I think he is correct that when coupled with laser-hologram technique, his technology can scale quite well. I do not understand the emphasis on metal bead since the laser manipulation part only require materials with asymmetric dielectric tensor. My guess is that magnetic manipulation is somehow involved in the construction of the library via a separate technique, like micro-fluidic to control fluid flow given the founder's background, or some fancy metal dna chemistry which i doubt because i do not think the founder like chemistry very much given his affinity. In short, he is building a fancy super-micro-sized dna library with super-dense micro-dna-arrays using light and magnet as micro-manipulators. He is definitely not 'printing dna' and definitely not developing a new sequence building method since that will require chemistry no matter what. Good marketing gimmick though.

  2. "I'm really not interested in discussing that, since I don't think there's much value to be added from me/us to the discussion."

    Aww, c'mon, how can this not be of value:

  3. This confused me for a bit too, which lead me here. From his interviews it seemed Cambria was synthesizing entire chromosomes or something. After looking through the patents it seems the newness here is just the sorting of many strands of DNA, sequence verified by some fluorescence method. It then seems like lasers sort all the correct strands into one tube and use PCR to assemble whole/parts of genes. Essentially just automated gene assembly? Patents are always hard to read so I could be wrong about that.

  4. It looks like he committed suicide last week. RIP. Nothing is worth that.

    1. Was it via synthetic DNA, or is this a setup perpetrated by BIG-DNA? Is this a PR stunt?

  5. My guess is that either the board tried to get rid of him as CEO, or he committed some sort of fraud related to the 10M investment 6 months prior. I have always been skeptical about how this oligo pooling technique stacks up against other technologies, but you don't give up like that after receiving 10M 6 months prior. There is just no way.