Coding 4 DNA

Play Coding 4 DNA
Sign in to queue


Andrew Phillips holds the title of Scientist with Microsoft Research Cambridge, and he's working on a method of programming that compiles into DNA. Part of this involves a visual programming language called Stochastic Pi Machine, or SPiM. This system models biological processes to help give researchers feedback on how organisms will react to modifications.

The hope is that this can be used to help scientists program for large biological systems using modular components compiled to DNA. Yes, I’m in way over my head here, but I do my best to ask Andrew about the role this will play in medical treatment going forward, what it means to DNA computing, and the ability of back-engineering the genetic code we don’t use now.



Right click to download this episode

The Discussion

  • User profile image

    Great research Andrew! I just read your paper introducing this concept. 


    It struck me that an interesting application of this would be to develop a treatment (not a cure) for cancer by maintaining cancerous cells below a certain clinically deleterious threshold.


    I would suppose the general algorithm would go something like:

    1) Select appropriate retrovirus as your output for your code that targets white blood cells (HIV based).  Have the "compiled" retrovirus spread through body.

    2) Retrovirus should code for expressing a weak binding receptor on the white blood cell that would target the cell type susceptible to cancer


    Normal scenario: If patient is healthy, the concentration of the target cell type are sufficiently low that the weak binding receptor doesn't impact the normal steady state concentration of the target cell type


    3) Cancerous scenario:

     - As target cell type concentration begins to increase, the weak binding receptor is occasionally activated

    4) Activation ofthe weak binding receptor then triggers a "positive feedback" reaction, where a strong binding receptor is now expressed.

    5) Strong binding receptor enables white blood cells to bind to target cell types (both healthy and cancerous, but by probability it will mostly be targeting cancerous) and destroy said cells.

    6) As concentration of cancerous cells falls, weak binding receptor is now inactivated (it only substantially activates above a threshold).

    7) Inactivation of weak binding receptor also inactivates the strong binding receptor.

    8) Rinse and repeat if cancer cells continue uncontrolled growth.


    Result:  Patient still has the cancerous cells, but you don't end up with runaway growth, and the steady state of the target cell type never reaches a clinically mailicious level.


    What do you think?  I'd love to work on a problem like this...


    -Deverraux J

  • User profile image

    Very interesting.  Photosynthesizing humans.  'Nuff said.

Add Your 2 Cents