anyone looked at long-range patterns in DNA sequence?

Ben Jones jonesbb at
Tue Mar 23 06:03:53 EST 1993

James Fickett writes:

>The Isochore model was, biologically speaking, rather mysterious:  what
>biological constraints would cause the C+G content to be held constant over
>megabase regions?  On the other hand the WM model is very natural.  The
>imperfect persistence of C+G content may be, in a naively simple analogy,
>similar to the clumps of related organisms one sees in any naturally
>occurring ecosystem.

(This message is a cleaned up version of one I sent privately to Mike
Levin.  I am posting it here because James Fickett has made it more
relevant to the net discussion thread.)

I hope nobody minds if I speculate wildly on this topic, but ideas from
different parts of biology suddenly collided somewhere behind my eyelids
and I didn't have time to duck.  (I get into more embarassing situations
that way. )

In a nutshell, the idea is that three-dimensional regions of the nucleus
might be high or low in CG base pairs, not just stretches of the
one-dimension of the chromosome.

As I recall from my undergraduate days, chromosomes in their normal state
are not just floating randomly in the nucleus.  (If they were, what a royal
mess it would be to untangle them for mitosis.)  They are rather fixed in
some kind of matrix, possibly by being attached to the nuclear membrane.  

Just suppose that the chromosomes are often (or always) attached to the
nuclear membrane in the same pattern in the nucleus.  I have no evidence
for this, but it seems possible.  If that is the case, then a given stretch
of chromosome might often be in proximity to a given stretch on adjacent
chromosomes, or a distant (by map distance) part of the same chromosome. 
In that case, then certain types of mutation might be more likely than
other types.  Perhaps there is some kind of thermodynamic advantage to
having CG in a region of the nucleus where there is already a concentration
of CG among all the strands that pass through a given neighborhood in the
nucleus, reagrdless of which chromosome they are on.

Here is a speculation about a mechanism.  An insertion mutation of a G or a
C may be simply more likely just because there are more CG base pairs being
transcribed, or copied, or repaired nearby.  (Or LESS likely, because a
neighboring strand might deplete the concentration of precursor GTP or CTP
in the area?)  (However, the diffusion distances involved would seem to be
extremely short in a nucleus, so perhaps a local concentration would exist
for only a vanishingly short time.  On the other hand, a DNA matrix may
slow diffusion down considerably.)

To test this idea, I'd love to see some electron micrographs of multiple
DNA probes in many different intact nuclei, to see if they show any
consistency in their three-dimensional spacial relationships.  This would
indicate that they generally occupy the same part of the nuclear space. 
Another test would be somehow to map the concentration of CG base pairs in
the nucleus to see if CG-rich isochores on different strands are physically
close to each other.  (Mapping the genome would take on a new meaning.  :-)


Ben Jones                  BioQUEST / Department of Biology
jonesbb at         Beloit College, Beloit, Wisconsin

More information about the Comp-bio mailing list