Go back to the main page.
Go back to the "extra" page.

The biotech toolkit project and the prospects of the bioreactor

Bryan Bishop, 2008-06-13

Author's introductory note. There are a number of DIY (do-it-yourself) free and open source (or shareable and public) projects accessible through the internet including the biotech toolkit, OpenWetWare, diybio, biopunk, biogang, and so on. Because of modern software architecture and technologies, there's the availability to use revision history systems such as git to proliferate collaborative community effort in the same way that the free and open source software (F/OSS) communities have successfully demonstrated (see GNU and Wp's article on 'open source'). In the following document, a simple system of user-oriented software, policies and physical technologies will be presented that emphasizes design strategies and minimizes wasted effort. Additionally, when one reads "community", this means any particular community, since everything can be copied and effortlessly applied in new contexts, thereby allowing anybody -- or any institution (societies, governments, regulatory committees, etc.) -- to have their say.

Anybody who has kept up with modern biology will have noticed bioinformatics. What's interesting about bioinformatics is that it is the application of computational methods, including databases and programming, towards the investigations within biological science research institutions. It allows ridiculously large datasets to be managed with automated machinery, both in terms of robotic arms to fetch yeast knockout libraries and in terms of processes running through databases to find interesting information, like possible cures to diseases, the similarities between different genomes, understanding ecologies or ecosystems and the development of seemingly innumerable advancements in both engineering and science. Although a great deal of these databases and programs are available to the public, what "grounding" is there? In other words, there is a limited availability of tools and machinery that people could use to do interesting things with, as long as their are the ridiculously high barriers to entry to set up labs and other things that, traditionally, require funding to get grants for and such.

The idea of do-it-yourself biotech kits and engineering design is that we can get rid of that overhead with a "bootstrap-once, use-many" system. The idea of the biotech toolkit project ( http://biohack.sf.net/ ) is to make a resource of organized, computationally-tractable (but also, most importantly, human-readable) information on how to implement this social knowledge that has been accumulating in scientific journals and other paper repositories for the past 40 years of computer-usage. What this allows is that through simple collaborative frameworks like git ( see gitweb.cgi or /biotech.git), individuals can contribute to a larger project and include valuable information that would otherwise be lost to obscure dead-tree format books and such. The difference between this and a wiki is small, but an important one. A wiki can be implemented on top of git (like ikiwiki or dokuwiki). However, the majority of wikis out on the web today use something called revision control or history, and usually this is implemented within the wiki software itself -- meaning that the development history of git, svn, cvs, monotone, etc., are completely ignored in place of custom implementations that are not nearly as robust and usually require database projects (MySQL, PostgreSQL, etc.). What this means for end-users, people clicking around and adding information, means little except that the usable set of functional possibilities is expanded rather than restricted by conventional wiki software.

Having information on protocols and materials is an interesting first step, but it is certainly missing the broader picture of what could be possible. This is where the bioreactor project steps in. Simply put, the bioreactor is a biologically "mostly" self-replicable kit that makes all of the materials one would need in order to modify the organisms involved in the bioreactor itself. The reason why it is 'mostly' self-replicating is because of the requirement of a chasis, which will probably be metal, plastic, or some other material to encapsulate different tanks or different cell cultures and so on. The bioreactor is being designed to incorporate all materials necessary, into the genomes of the organisms involved. Friends would be able to make a new bioreactor for another friend, and a chain of proliferation and exponential growth could be developed. Individuals could use it to test out new ideas, to try out do-it-yourself biogerontology experiments, to make biomolecular compounds and run them through purification processes, etc. There are many, many possible uses of such a bioreactor.

The main component of this bioreactor is the idea of a DNA synthesizer incorporated into the biology itself, perhaps an in vitro DNA synthesizer, hopefully in vivo. This component (still a subject of active R&D) would allow new strands of DNA to be implemented on the spot without the use of bulky DNA synthesizers using silicon manufacturing that, frankly, not everyone has access to. The in vitro DNA synthesizer (writozyme) may require either a "retarded" polymerase or protein-based oligonucleotide synthesis. All of the development patterns of the biotech toolkit, the repositories, books, information, communities and so on effortlessly map over to the bioreactor project, with all of the benefits and more.

Resources on git

Git - fast revision control system (asynchronous, amorphous, distributed development)
http://git.or.cz/gitwiki/InterfacesFrontendsAndTools (for graphical user interfaces)

Google TechTalk on git or see below:

Here's a video on the toolkit and git.

There's also an IRC channel (#hplusroadmap on freenode).

Maybe you'd like to see the wiki?

The page's predecessor?. It's not quite the predecessor, for manufacturing is still the central focus, however biology provides some useful pieces that can be used to kickstart a number of interesting projects -- especially bioinformatics, which has excellent protocols and support that manufacturing, sadly, simply can't match at the moment (and thus sucks for the basis of a description of what I'm talking about).