Additional Paper: Transforming Cells into Automata

designpadΤεχνίτη Νοημοσύνη και Ρομποτική

1 Δεκ 2013 (πριν από 3 χρόνια και 8 μήνες)

113 εμφανίσεις

CS 374: Algorith
ms in Biology



Fall 2004

John Shedletsky

Additional Paper: Transforming Cells into Automata


Paper reference



Batten C, Krashinsky R, Knight T. “A Scalable Cellular Logic Technology Using Zinc
-
Finger
Proteins”. MIT Computer Science and Artificial Intelligence Laboratory, Cambridge, MA 02139.
htt
p://www.cag.csail.mit.edu/~ronny/papers/zfp
-
engbio.pdf


Abstract



The technique of transforming a cell into an automaton capable of processing simple logic routines,
whil
e still in its infancy, has several unique benefits and holds great promise. Some of
these advantages
include the fact that such a computational substrate would be, as a matter of course, tightly integrated with
biological inputs and outputs. Cellular computation engines would also be easy to grow in large numbers
and could possibly be tun
ed using the process of directed evolution. On the other hand, significant
engineering obstacles must be overcome before this promise can be fully realized. The first is that the cycle
time for cellular automata can range on the order of hours. The second
is that only modestly complicated
circuits have been realized thus far

as a result of the limited number of repressor proteins exist, targeting a
limited number of unique binding regions on a DNA strand. Since these repressors compose the gates of any
circ
uit one might attempt to execute in a cellular automata, this constitutes a serious limitation.

This pap
er proposes a technique to help alleviate the problem of there not being sufficient known
natural repressor proteins for use in creating complicated pr
o
karyotic cellular automata. The innovation is to
use synthetic zinc
-
finger proteins (ZFPs) to create a vast library of repressor proteins that can bind to an
arbitrary sequence of base pairs on a DNA strand. At present, it has been demonstrated that it is

possible to
engineer a ZFP to bind to any “G**G” or “A**A” sequence. It has also been demonstrated that single ZFPs
can be concatenated to create poly
-
finger ZFPs, which can recognize longer series of base pairs.

To increase
the effectiveness of the ZFPs
in binding to DNA, the researchers fused a pair of two
-
finger ZFPs to a GCN4
leucine zipper dimerization. The resultant molecule can recognize a unique 6 base pair binding site and has a
binding affinity suitable for biological circuit dynamics. The scala
bility of these dimerized gate components
has not been well explored, but it
is
estimated that circuits with several hundred gates will be realizable.
The
researchers state that science is nearing the point where an appropriate poly
-
finger ZFP can be easil
y
composed from a library of fingers to recognize almost any DNA sequence


Discussion





This paper, published very recently, represents significant progress in the field of transforming cells
into automata. From the research presented in class, previou
s experiments have demonstrated simple logic
circuits, the most complicated being a toggle switch, requiring only a handful of repressor proteins to
implement.
Other current systems seem to be

limited to less than a dozen gates


Three inverter ring
oscill
ator [Elowitz00], RS Latch [Gardner00], Inter
-
cell communication [Weiss01]
. While the researchers in
this paper did not actually construct a larger cellular automaton, they present a plausible technique that could
be used to do so. The current 2
-
ZPF
-
GCN4 f
usion dimer
introduced in this paper is specific enough to allow
the creation of logic circuits in a prokaryotic genome. Perhaps more interestingly, the researchers have done
experiments which demonstrate heterodimer fusions have been made with poly
-
finger

ZPFs that can
recognize unique 12
-
18 base pair sequences


providing sufficient specificity to perhaps allow the creation
of cellular automata in eukaryotic cells (where the genomes are considerably larger and more complex). This
is an important step that

must be taken if cellular automata are going to be anything more than a laboratory
tool
(eventual applications could include therapies that intelligently target diseases).


An area of current research in ZFP repressors is the problem of reducing inter
-
gat
e interference, as it
relates to scalability and gate delay.
The authors suggest that future research should focus on increasing the
chemical cooperativity of the system. High cooperativity would enable faster and more robust logic gates.