Re: [DIYbio] Advice for a computer scientist who doesn't understand biology

Ujjwal,
Here's another discussion that started off talking about the term
thrown around by a lot of new DNA design softwares 'github for
biology'... and how you could simply just use git already, so it's
nothing too new. The discussion then morphs into ASCII vs other types
of digital DNA representations, considering space required and
additional features that DNA can posses other than A T G or C (i.e.
methylation). The discussion really get's interesting when Anselm
said:

On Fri, Sep 7, 2012 at 2:48 PM, Anselm Levskaya <levskaya@gmail.com> wrote:
> There's too much comparison going on here with von neumann machine code.
>
> DNA "code" is intrinsically nonlocal, i.e.:
> - when you're editing a protein-coding ORF, there are long-range
> correlations across the entire structure of the gene that corresponds
> to the 3D interactions that define the shape and dynamics of the
> protein.
> - when you're playing around with regulatory elements (cis factors)
> and their protein regulators (trans factors), you're also dealing with
> a network of interactions that are non-local and complexly
> intertwingled across the sequence.
> - eukaryotic cells add a whole new level of state-machine logic on top
> of local (de)compaction as mediated by chromatin remodelling systems
> and poorly-understood insulator/enhancer systems.
>
> So you can't just expect to write useful linear source code unless
> you're doing the most primitive cut-and-paste style synbio of a
> handful of unmodified genes. i.e. just screwing around. (which is fun
> and great - but not really what synbio is ultimately about)
>
> Abstractions work in machine code because we built the machines to
> make abstractions possible. Natural cells don't work like that. The
> closest thing to cells in the machine code world are demoscene x86
> assembly blobs -- filled with insane hacks to make something awesome
> work in a small amount of space, with lots of weird code and data
> reuse and generative magic.
>
> It's worse than that though. Never forget that the true "compiler" of
> DNA is Physics, specifically that of protein folding, conformational
> dynamics, and catalysis. It won't be simulated anytime soon with
> anything approaching useful kinetic accuracy. (It's not clear if
> we'll ever get kT-accurate quantum simulations of correlated electron
> wavefunctions that scale to protein-sized systems, though there is
> some hope in the far future with exotic computing architectures.)
>
> So ultimately what software useful for synthetic biology is going to
> do is help us curate all of our brute-force efforts to build and
> screen libraries of pseudorational libraries of proteins, pathways,
> and cells. -Not- provide shitty abstraction layers that rest upon our
> incredibly shaky understanding of what's going on circa 2012. Think
> of curating an incredibly complicated genetic-programming run across a
> hundred-thousand clusters -- that gives a better sense of the flavor
> of what's needed.
>
> We invented the light-build long before we understood QFT. We'll be
> building amazing cellular machines long before we really understand
> them quantitatively. Synbio's (and diybio's) biggest sin is
> repeatedly elevating the convenient metaphors with EE/CS into a
> slick-looking action plan that doesn't respect the fundamental
> differences between these machine architectures.
>
> -a

Main discussion
https://groups.google.com/d/topic/diybio/GxRTESzUWUI/discussion


--
-Nathan

--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diybio@googlegroups.com. To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diybio+unsubscribe@googlegroups.com.
To post to this group, send email to diybio@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.