-----------------------------------------------------------------------
		       BIOINFORMATICS COLLOQUIUM
		    School of Computational Sciences
			George Mason University
-----------------------------------------------------------------------
 Shifting the Balance between Physical Experimentation and Computation

			  Richard J. Feldmann
		       Global Determinants, Inc.
		    rjfeldma@GlobalDeterminants.com

		       Tuesday, September 9, 2003
				4:30 pm
	     Room 110H, Building PW1, Prince William Campus

Something rather amazing happened in May at the first day of the NIH
symposium honoring the 50th Anniversary of the publication in Nature of
the Watson and Crick paper and the completion of the sequencing of the
human genome.  The speakers who were chock-ablock about their work on
the human genome, one after another argued that computation would now
replace physical experimentation as the driving force in biological
science.  There was one speaker who talked about his work to identify
the genes in the human genome and to throw out the pseudo-genes.  A
couple of weeks later Hirotsune et al. published a paper showing that
the deletion of a pseudo-gene affected the expression of its cognate
gene P so now pseudo-genes clearly have function.  Eventually one might
expect to show function for some parts of the non-genic 97% of the human
genome.

Fours years ago while thinking about the design of zinc-finger DNA
Binding Proteins (DBPs), I stumbled on the existence of tetrads of DNA
sequences that act to control gene expression.  I tested the first
hypothesis in yeast and found that there were reasonable numbers of
these tetradic relationships that I now call Connectrons.  I showed that
some genes have two signals in their 3UUTR that when expressed as RNA
move about in the whole cell of prokaryotes and Archea and the nucleus
of eukaryotes.  These RNA signals bind to two equivalent DNA
double-stranded sequences to form generalized, triple-stranded Hoogsteen
helices.  Since the RNA signals are covalently linked, the two DNA sites
are brought together in space.  This allows the loop of up to 100kb to
be structured in six-fold-symmetric 30-nm chromatin particles.  The
genes that are in the chromatin-structured loop are not available for
promotion and expression.  Of course, the connectrons have life-times
that are proportional to the length of the interaction between the RNA
and the DNA, so after a while a connectron will open up and the genes in
the loop will be available once again for promotion and expresssion.

I have gone on to show that all the genomes on the NCBI site have
connectrons.  I will tell you all about connectrons and give you a CD to
look at later.

I have just finished detrmining the connectron structure of the human
genome and I am using my cluster of PCs to do the mouse genome now.

My computation of connectron structure is what other people call
data-minimg.  I have tried to form collaborations that would lead to
physical proof of the existence of connectrons but so-far these
experiments have failed.

So, where does the balance between computation and physical
experimentation stand now?  I would argue that by providing a
whole-genome approach to gene expression regulation, I have shifted the
balance in favor of computation.  It's just that nobody else seems to
believe me.  You can decide for yourself.  The most conservative will
cling to "one-gene / one-effect" while the adventuresome will jump to
thinking about connectrons.

----------------------------------------------------------------------
Refreshments are served at 3:30 pm.  Find the schedule and directions at
http://www.binf.gmu.edu/colloq.html