DNA MESSAGES
Cliff Pickover
and biologists, so serious readers will wish to skip this posting. If
you are interested, I'd be curious as to your answers to the few
questions which are highlighted in CAPITAL letters. I am interested in
the coding and information aspects of these questions. You may send
your ideas and comments directly to me, if you wish. This might
avoid cluttering the bulletin board with postings which some
readers may wish to avoid due to their speculative nature.
cliff @ watson.ibm.com
DNA MESSAGES
It's easy to imagine a science-fiction plot where
a scientists sequencing the hereditary material of life (DNA) notes that
it encodes a message. The DNA sequence, expressed as a four letter code
using the symbols G, C, A, and T, could code for practically anything.
For example, it could contain the value for pi (3.1415...). HOW MANY
DECIMAL DIGITS OF PI ENCODED IN DNA WOULD IMPRESS A SCIENTIST AS
SOMETHING SPECIAL? After all, just a few digits could occur simply by
chance. A few hundred digits might make a scientist gasp in shock.
Astronomers for years have scanned the heavens for radio messages
from alien creatures. IF YOU WERE AN ALIEN CREATURE TRYING TO CODE A
MESSAGE USING THE FOUR SYMBOLS (G, C, A, AND T), HOW WOULD YOU DO IT,
AND WHAT MESSAGE WOULD YOU ENCODE?
The idea of placing messages in genetic sequence is not
entirely fanciful. Joe Davis at MIT hopes to place encoded messages in
the DNA of a bacteria which could duplicate and spread through the
galaxy. His collaborator, Dana Boyd, a geneticist from Harvard, has
synthesized a short sequence of DNA consisting of 47 base pairs with a
brief coded message. When converted to a grid of binary digits, the
message appears as a sketch of the female genitalia. 100 million copies
of this message have been stored in a vial. Of course Davis and
colleagues do not really plan to disperse these bacterial spores, but
Davis has noted that this "may be the only practical way for humans to
explore the cosmos."
USING JUST THE 4 SYMBOLS G, C, A, AND T, WHAT ARE SOME WAYS
YOU CAN THINK OF FOR CODING MESSAGES FOR TRANSMISSION INTO OUTERSPACE?
POST...@gunbrf.bitnet
The premise is, in fact, similar to that of the science fiction novel
"Andromeda Strain".
With DNA's 2 bits, unpunctuated, per base, an encoding of PI might proceed
in this way:
(1) establish order and frame
T C A G T T C A G T T C A G T
(2) begin message
G T A C T C ... (to five decimal places)
A non-repeating message, like PI, could easily be mistaken for noise
(complexity theory). An more easily recognizable message would repeat or
otherwise exhibit patterns.
T C A G T T C A G T T C A G T T T T T T T T T C T C T C T T T C T
T T A T A T A T T T A T T T G T G T G T T T G T C T C T A T
C T A T G T A T C T G T C T G T C T T G T C T C T T ...
I believe some Ozma, Explorer and Voyager messages have tried to encode bit-map
pictures based on prime number sized arrays. These were simple 2-dimensional
arrays that would hardly show up on a low-resolution monitor. Today they would
probably make 439 x 1021 pictures. But why not 3-D arrays showing something
like a human body, or 4-D and let it wave its hand.
A more informative message would perhaps assume that a compatible biochemical
soup existed at the destination (that is, that the DNA would be translated to
RNA) and would set out the plans for an RNA machine that, perhaps among other
things, went about reproducing the message. It would look like a virus with no
protein dependence. At this point I will defer to Tom Schneider for the
construction of such a molecular machine.
------------------------------------------------------------------------
Dr. John S. Garavelli
Database Coordinator
Protein Identification Resource
National Biomedical Research Foundation
Washington, DC 20007
POSTM...@GUNBRF.BITNET