Biological manufacturing

6 views
Skip to first unread message

Bryan Bishop

unread,
Sep 15, 2008, 8:18:35 AM9/15/08
to openmanu...@googlegroups.com
All,

I've neglected to mention some of my biologically inclined manufacturing
projects. And since I have to run in a few moments, I'm going to leave
a few links and short descriptions. I was hoping to use these projects
as something to get rolling /before/ the hardware manufacturing blasted
off. The mention of composting reminded me to mention their existence:

Synthetic biology toolkit / biotech toolkit:
http://biohack.sf.net/
http://heybryan.org/biotech.git
http://heybryan.org/mediawiki/

The toolkit exists to make biological projects easier for individuals.
It's still in need of some work, mind you. But it does have enough
information to get started on making DNA synthesizers, sequencers,
microarrays, or doing kitchen-based experiments.

Bioreactor project
http://heybryan.org/new_exp.html
http://heybryan.org/~bbishop/docs/bioreactor/

The bioreactor is my extension to the toolkit. Imagine a giant tank that
produces all of the reagents and molecules necessary to pull off the
otherwise highly expensive experiments in the biotech toolkit. Since
it's biologically based, this entire 'tank' is self-replicable. One of
the major aspects is the in vitro biologically based DNA synthesizing
method that I'm still working on.

These factors may be useful in some more manufacturing options. Any
takers?

- Bryan
________________________________________
http://heybryan.org/
Engineers: http://heybryan.org/exp.html
irc.freenode.net #hplusroadmap

Vinay Gupta

unread,
Sep 15, 2008, 8:33:57 AM9/15/08
to openmanu...@googlegroups.com

(reading)

This looks fabulous, very, very interesting stuff. I'm not a bio guy
(at all) although I'm beginning to learn a bit of this stuff from
Paul Stamets work, so... interesting stuff, interesting stuff indeed.

Vinay


--
Vinay Gupta
Free Science and Engineering in the Global Public Interest

http://hexayurt.com - free/open next generation human sheltering
http://hexayurt.com/plan - the whole systems, big picture vision

Gizmo Project VOIP : (USA) 775-743-1851
Skype/Gizmo/Gtalk : hexayurt
Icelandic Cell : (+354) 869-4605

"If it doesn't fit, force it."

Josef Davies-Coates

unread,
Sep 15, 2008, 9:10:21 AM9/15/08
to openmanu...@googlegroups.com
2008/9/15 Vinay Gupta <hexa...@gmail.com>

(reading)

This looks fabulous, very, very interesting stuff. I'm not a bio guy
(at all) although I'm beginning to learn a bit of this stuff from
Paul Stamets work, so... interesting stuff, interesting stuff indeed.


Glad you mentioned Paul, was just about to do so myself :)

The stuff he is doing with Mushrooms is amazing.

His TED talk here:
http://www.ted.com/index.php/talks/paul_stamets_on_6_ways_mushrooms_can_save_the_world.html

I think I'm going to buy his book...

Josef.

--
Josef Davies-Coates
07974 88 88 95
http://uniteddiversity.com
Together We Have Everything

Smári McCarthy

unread,
Sep 15, 2008, 9:13:55 PM9/15/08
to openmanu...@googlegroups.com
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Interesting stuff. Signed for mail. I'll definitely have to take a good
look at this when I have a chance.

I'm personally most interested in functional programming using
proteins.. consider proteins being built to synthesize plastics in a
vat, for example. It's not as far-fetched as one might think. My nascent
theory about protein-protogenesis has also been amazingly well received
by those in the know... wish I had time to do a proper rewrite of the paper.

In this vein I thought I'd also point out the Center For Bits and Atoms'
Millibiology project, sponsored by DARPA. There's not a lot of info on
the Internet besides references to the project's GIT archive (no
surprises actually...). What little more I know I'm not sure I'm at
liberty to share, but essentially the goal is to engineer
implementations of cellular functions on a micrometer/millimeter scale.
Hence the name.

- Smári
- --
Smári McCarthy
sm...@yaxic.org http://smari.yaxic.org
(+354) 662 2701 - "Technology is about people"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFIzwhT9cJSn8kDvvERAjmCAKC0eMi3UmwxgvDv5pGDUWWqDIjgqQCfdq2C
7F76H/06mUvYkDmcTODkM1U=
=yOlS
-----END PGP SIGNATURE-----

Bryan Bishop

unread,
Sep 15, 2008, 10:49:32 PM9/15/08
to openmanu...@googlegroups.com
On Monday 15 September 2008, Smári McCarthy wrote:
> I'm personally most interested in functional programming using
> proteins.. consider proteins being built to synthesize plastics in a
> vat, for example. It's not as far-fetched as one might think. My
> nascent theory about protein-protogenesis has also been amazingly
> well received by those in the know... wish I had time to do a proper
> rewrite of the paper.

You have to be careful because we don't have the ability to engineer
proteins from scratch yet. So one of the projects I mentioned in my
first email on this mailing list was proteinCAD, or a synthetic biology
circuit creator app. This 'circuit creator' is proceeding smoothly, I
have data sets on the server that you can play with, but there's a big
problem. Ask one user for what sort of input he would imagine into such
a system, and then go ask another person for their input as well.
You'll find that the format for input of specifications of a created
circuit, is a terribly complex problem. I'd appreciate any insights
into this .. :-)

> In this vein I thought I'd also point out the Center For Bits and
> Atoms' Millibiology project, sponsored by DARPA. There's not a lot of
> info on the Internet besides references to the project's GIT archive
> (no surprises actually...). What little more I know I'm not sure I'm
> at liberty to share, but essentially the goal is to engineer
> implementations of cellular functions on a micrometer/millimeter
> scale. Hence the name.

Yeah, I talk with CBA every day. This is one of their projects I wasn't
aware of, however.

Helen Titchen Beeth

unread,
Sep 16, 2008, 1:57:06 AM9/16/08
to openmanu...@googlegroups.com

On 16 Sep 2008, at 03:13, Smári McCarthy wrote:

> wish I had time to do a proper rewrite of the paper.

Smári, If this paper is as important as you intuit that it is, then
what would stop you from putting the time into rewriting it?

:-D

h

Smári McCarthy

unread,
Sep 16, 2008, 7:05:34 AM9/16/08
to openmanu...@googlegroups.com
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Helen Titchen Beeth wrote:
> Smári, If this paper is as important as you intuit that it is, then
> what would stop you from putting the time into rewriting it?

Mostly that my background is mathematics, not biochemistry, meaning I'd
need to put several months of solid work into it at minimum, and between
trying to run a Fab Lab, soon possibly two, teaching courses in two
places 100 km apart, taking a course at a third place several thousand
kilometers away, doing research, development and documentation,
pretending to be writing two novels, actually reading several, and all
the other madness I've managed to shanghai myself into, I kind of don't
have several months to spare right now. The time will come, don't worry,
but for now I'm fairly set. :)

Bryan Bishop wrote:
> You have to be careful because we don't have the ability to engineer
> proteins from scratch yet. So one of the projects I mentioned in my
> first email on this mailing list was proteinCAD, or a synthetic biology
> circuit creator app. This 'circuit creator' is proceeding smoothly, I
> have data sets on the server that you can play with, but there's a big
> problem. Ask one user for what sort of input he would imagine into such
> a system, and then go ask another person for their input as well.
> You'll find that the format for input of specifications of a created
> circuit, is a terribly complex problem. I'd appreciate any insights
> into this .. :-)

You may or may not well be my new hero. Probably not, but I'll
definitely give you feedback.

Intuitively I can imagine, without having seen your app, a text editor
that understands three letter amino acid codes on the one hand and
chemical formulas in SMILES format on the other. That way syntax could
be handled like:

Trp-Ala-Met-Phe-Pro-Gln-Ser-Cys-Trp-Asp-Asn-Met

Or:

Trp-<C[C@H](N)C(O)=O>-Met-Phe-Pro-Gln-Ser-Cys-Trp-Asp-Asn-Met

The former would be far easier to simulate the behaviour of unless you
add something that recognizes C[C@H](N)C(O)=O as Alanine, but for
non-standard amino acids like, say, selenocysteine,
N[C@@H](C[SeH])C(O)=O, you'd have to have either a very detailed
database of amino acids and their behaviours or a very very good system
for predicting their behavior, which is probably not horrendously
difficult to do if you have ten grad students and a decade to develop
it. Do you?

The text input format would then be coupled to a graphical
representation where you could click-and-drag to manipulate protein
structure on a per-amino acid basis.. then you could, say, select an
amino acid and explode it to mess with it's internals. Likewise you
could group a bunch together and zoom out of them to work with larger
bulks, since a lot of the work you'd do with such a program would be
with tertiary/quaternary structures as well as drop-in functionalities
like, say, protein segments like enzimes (say, hemoglobin or insuline)
or motors (say myosin), or whatever.

In short: Graphical format that provides for multiple scales of
interaction, and an exposed internal format view that allows direct
manipulation of the protein by way of a text editor.

But that's just me.

> Yeah, I talk with CBA every day. This is one of their projects I wasn't
> aware of, however.

As do I. The world truly is small.

>
> - Bryan

- Smári


- --
Smári McCarthy
sm...@yaxic.org http://smari.yaxic.org
(+354) 662 2701 - "Technology is about people"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFIz5L99cJSn8kDvvERAnVvAJwOVUWfJmRCAj83i5/dlcCDuDYJLACdHx8F
3Lj0vASKObaV987IFNurPT8=
=E/MZ
-----END PGP SIGNATURE-----

Bryan Bishop

unread,
Sep 16, 2008, 8:06:03 AM9/16/08
to openmanu...@googlegroups.com
On Tuesday 16 September 2008, Smári McCarthy wrote:
>   In short: Graphical format that provides for multiple scales of
> interaction, and an exposed internal format view that allows direct
> manipulation of the protein by way of a text editor.

Hah! You're going to be blown away by this one. The input is going to be
a bit more high level than that. Nobody ever sees the DNA or amino
acids, unless they absolutely have to. It's built up from biobricks and
other related parts (so I don't need 10 grad students (who am I
kidding, I'd love 10 grad students!)) that are pulled from the
bioinformatics databases. Sort of - there's a few tricks I'm employing
here that we get to handwave - ignore the handwaving for the moment.

What I'm actually wondering about, though, is the problem that I can
assemble any circuit possible with all of these parts. Remember, I'm in
the Automated Design Lab. Screw human interaction. ;-) Okay, I'm lying,
I want to include human interaction, but the problem still persists
even when the computer is doing the designing. When you first sit down
at the computer to design your protein, or to design your genetic
regulatory netwrok, what do you *do*? "Computer! Make me a system that
is able to secrete ozone!" Most people don't talk or think like this.
Rather they are going to be talking crpytically and
wanting 'specifications' that are, frankly, computationally hard to
define or evaluate. Another equivalent example is when you're
consulting for somebody when you're writing software. They give you the
most ridiculously useless set of specs and requirements, and you can't
really tell them "No, you're an idiot" because you don't want to be
punched.

One method that an MIT student used once was to "repair" the user
specifications to translate it from the semantic la la land of the user
into the parts and components in the database that the computer knows.
Thus intuitively starting with the theorem that the user is wrong. It's
fancy, but it doesn't solve the entire problem.

Smári McCarthy

unread,
Sep 16, 2008, 7:13:55 PM9/16/08
to openmanu...@googlegroups.com
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Fair enough, but remember that "high level languages" as you're
describing are reliant on low level languages. What you're describing is
the "python" or "perl" of protein design software, while what we need at
the moment is the "C" of such software, with a clear method of looking
beyond into the assembler code. Hence my focus on the amino acids and
the SMILES. Being able to plug entire proteinical engines to each other
is smart and cool, but at the moment it's too large grained to be what
we really need.

Have cloned your GIT both at home and at work, and will dredge through
the files when I have a chance.

- Smári


Bryan Bishop wrote:
- --
Smári McCarthy
sm...@yaxic.org http://smari.yaxic.org
(+354) 662 2701 - "Technology is about people"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFI0D2y9cJSn8kDvvERAp55AJ9PgxPkdx0AUTNgSL4kKDrnF11KogCfafED
4dKEB+daGGEGXpx/SGMIph8=
=uXh9
-----END PGP SIGNATURE-----

Bryan Bishop

unread,
Sep 16, 2008, 9:32:15 PM9/16/08
to openmanu...@googlegroups.com
On Tuesday 16 September 2008, Smári McCarthy wrote:
> Fair enough, but remember that "high level languages" as you're
> describing are reliant on low level languages. What you're describing
> is the "python" or "perl" of protein design software, while what we
> need at the moment is the "C" of such software, with a clear method
> of looking beyond into the assembler code. Hence my focus on the
> amino acids and the SMILES. Being able to plug entire proteinical
> engines to each other is smart and cool, but at the moment it's too
> large grained to be what we really need.

No, you misunderstand me again :-). I'm telling you that that we do have
the low level stuff going on, to some extent I'm able to do that. It's
not as low level as exploiting the physics of protein folding blah blah
blah, sure, but it's low enough. It's not going to be protein folding
of course, it's just the utilization of protein information as we know
it. Protein folding could plug in eventually, of course.

Now. When you sit down at your computer to design a new synthetic
biology system. You could be considering many different things. The
first step that the user has to execute is, what exactly? I can imagine
many different projects that one might want to implement. Suppose you
want to make metal eating extremophiles that can survive the cold
realities of space. Suppose you want to make a cinammon-smelling ecoli.
Maybe you want to make bacteria that can display a two-color
photograph. All of the proteins and DNA for doing such already exists.
In fact, for all practical purposes it's already in my circuit creator
system. But. As the designer, you don't really know that because if you
did, you'd skip my designer all together and just get started :). Each
part in the database tells you things like the genetic inputs, outputs,
and to some extent some encoded functionality if you want to do some
major simulations for optimization and such for part-part interactions.
These inputs and outputs are variables in the forms of various
scientific units for measurement. Rate constants, volume, density,
various stuff like that to characterize each part. Mostly
stoichiometric information - so that you can see that 2 parts this, 4
parts that leads to 6 moles of the other. Not all problems reduce to
simple stoichiometry (do they?) so I'm specifically seeking an
alternative that users will be using as the handlebars to using the
system. Conceivably, they could just sit down and have a list of steps
in the synthetic biological circuit or something, and just choose
a 'general' part to include at that spot, and they find these 'general
parts' via searching functionality, knowing people, tutorials,
whatever; but people don't think like this .. they don't know the exact
way to begin their design. Let the computer do the beginning. You still
need to have the user tell the computer something, however. What is
this information? I'm not about to create some arbitrary ontology for
the construction of different types of projects .. i.e., bacteria that
secrete things would be one type, while bacteria that do chemotaxis
would be another; no, I don't want these silly separations. I'd prefer
not to, but perhaps I have to. How unfortunate. :-( (But still
awesome.)

Reply all
Reply to author
Forward
0 new messages