Below is an abstract for my general Spring 2000 talk. 
Mark Gerstein
--

Title:

Comparative Genomics: Surveys of a Finite Parts List 

Abstract: 

My talk will be deal with topics from comparative genomics, structural
genomics, and large-scale analysis of gene-expression data. I will
focus on how the "finite list of protein parts" can help simplify and
interpret genome sequences and expression data. I'll try to touch on
some of the specific points below:

1. A PARTS LIST OF FOLDS. An essential requirement for a structure
survey is a library of folds, which groups the known structures into
"fold families." I will describe various aspects of a fold library,
including methods of structural alignment and how important objective
statistical measures are for assessing similarities within the
library. I will also describe how a library of folds can be used to
precisely parameterize the degree to which structural annotation can
be transferred as a function of sequence divergence.

2. THE OCCURRENCE OF PARTS IN GENOMES. One can use a fold library to
count the number of folds in genomes, expressing the results in the
form of Venn diagrams, "top-10" lists, and whole-genome fold trees for
shared and common folds. One particular analysis shows that the common
folds shared between very different microorganisms - i.e. in different
kingdoms - have a remarkably similar super secondary structure, being
comprised of repeated strand-helix-strand units.

3. FUNCTIONS OF THE PARTS. Next, I will turn to relating structure to
function, looking comprehensively at how many functional roles a given
structural part can have. This will focus on the occurrence of folds
and functions in the yeast genome. Also, I will present a measure of
to what degree functional annotation can transferred at given amounts
of sequence identity.

4. USING PARTS TO ANALYZE EXPRESSION DATA. Finally, I will look at how
protein parts and their attributes can be used to help analyze
whole-genome expression data. I will relate look at the structural and
functional characteristics of the most highly expressed proteins in
the yeast transcriptome and develop a list of the most highly
expressed folds. I will show how simple clustering of microarray data
does not necessarily find proteins of similar function. I will also
show how expression level is correlated wtih the subcellular
localization of proteins and how this fact can be combined with
sequence patterns to predict localization in a Bayesian framework.

Continuously updated tables and further information pertinent to this
talk is available over the web at http://bioinfo.mbb.yale.edu/genome.

The talk is available from http://bioinfo.mbb.yale.edu/lectures,
sublink http://bioinfo.mbb.yale.edu/lectures/spring2000 .