DNA sequencing/info-theory paper in PNAS

[Steven Brenner]

In article <1993Mar23.220831.23581 at ncsu.edu>, samodena at csemail.cropsci.ncsu.edu (S. A. Modena) writes:

> In article <1olvdkINNbun at shelley.u.washington.edu> venk at stein.u.washington.edu (Venkatesh Murthy) writes:

>>Sequencing two DNA templates in five channels by digital compression.
>>Michael Nelson, Yanping Zhang, David L. Steffens, Reingard Grabherr
>>and James L. Van Etten.  PNAS 90:1647-1651 (March 1 1993).

... sounds interesting, I'll have to read it.

> This is not the way you want to do it IF a *human* is going to read the
> lanes off the autorad!  But it is extremely well suited to AUTOMATED
> sequencing where through put/cost trade-offs are critical in a truely
> massive sequencing project.

Hrmm.... I'm not so sure I agree with this statement ^^^.  Sequencing
gels have proven incredibly difficult to read automatically.  Under
the "compression" system, the some of the biggest hints for
gel-readers (i.e. the hints which make these things viable at all)
would be lost.  This really has nothing to do with biological
information theory, but I just wanted to do the sanity check on the
next comment:

> I might add a side comment or two about one of the "critical readers"
> acknowledged: Myron Brakke.  Brakke (if it's the same one  :^) ) is a plant
> pathologist who years ago simplified the preparative isolation/purification
> of plant virus particles from homogenized leaf material by *inventing*
> sucrose density gradient centrifugation.......and I recall from a
> conversation with Mathew Messelson how critical the timing of that
> invention was to him and Stahl (wasn't it) performing the "famous"
> experiment that UNequivically proved the conservative model of replication
> for ds-DNA...which if you recall, hinges on exactly the *same* decoding
> scheme for interpreting the results as I mentioned above for working
> through the two-lane sequencing example.  

That's SEMI-conservative replication, isn't it?  (Two new ds-DNA, each
of which has one strand half old, and one half new).  [The other
scientist was Stahl, I believe; the two met as a result of a practical
joke being played by James Watson.]

But how exactly does this relate to the decoding scheme?  As I recall,
in the Meselson-Stahl expermiment, cells were grown in heavy media
(N15?) for a long time, and then switched to light for a few cell
divisions.  The ultracentrifugation runs looked like this:

t=          0       1       2      3

Top                        ----  ====
(light)     

                   ----    ----  ----

Bottom
(dense)   ----


Indicating, of course, that at time zero, all of the DNA is
heavy-labled.  After one round of replication, all of the molecules
are half-heavy, half-light (due to having one of each type of strand).
At time 2, however, there are some all-light molecules, and some
middling ones.  By time 3, most of the molecules are all-light.

Maybe it relates to the time at which I'm reading, but I assume I'm
entirely missing the information-encoding point here.  Could you
explain it further?

-Steven

--
Steven E. Brenner               |  Internet    seb1005 at mbfs.bio.cam.ac.uk
Department of Biochemistry      |  JANET       seb1005 at uk.ac.cam.bio.mbfs
University of Cambridge         |  Laboratory  +44 223 333671
Tennis Court Road               |  Home        +44 223 314964
Cambridge CB2 1QW, UK           |  Lab Fax     +44 223 333345