Transcript PowerPoint slides

Classical tree view of cell cycle data
(Spellman, et al. 1998. MolBiolCell 9, 3273)
VxInsight topography of cell-cycle data
G1
M
S
Strong similarities show
relationships among clusters
How can we learn more
from this analysis?
Genome-scale datasets available in yeast:
•Essential genes
•Essential genes since 1998
•Several microarray datasets
•Protein-protein interactions
Essential genes as a function of gene expression:
what does it tell us?
Ribosome ridge
Stationary-phase genes are not
essential
Essential genes
Newly identified essential genes
Assumptions; biases; potentially new, useful targets, how cells
“protect” themselves; evolution, etc.
Comparison of gene-expression datasets
to test hypotheses
Are G1-regulated genes clustered during exit from stationary phase?
Cell cycle
Exit from stationary phase
What might this say?
Exit from stationary phase is either:
1) not a synchronous process with respect to the cell cycle.
2) a cell-cycle process that requires a subset of cell-cycle genes
Exit from stationary phase
•If this is true, the two processes may be
sensitive to different toxins.
•This has important implications in
treatment of infectious diseases
if the infectious agent spends a great deal
of time in the quiescent state
•This may also help us understand why
unculturable microorganisms can’t exit
stationary phase
Cell cycle
Protein -interaction networks as a function of gene expression
Schwikowski’s data
Interactions common to both
Ito full dataset
Conclusions:
1. Interactions do not generally
follow expression patterns
2. Few interactions common to both
datasets
3. Need to look at specific clusters,
known interactions to determine
whether one dataset should be
Genes common
to both
accepted
or whether
the data
should be combined
A Exit from stationary phase’
Ribosome ridge. 290 genes
B Similarity in gene expression
Conclusion:
Two-hybrid methods don’t “see”
interactions between ribosomal
proteins
C
Interactions in ribosome ridge
In fact, there may not be many
interactions among ribosomal
proteins – so this may be the
strongest evidence for the lack
of false positives in this analysis
Summary:
Visualization of the datasets enables a more intuitive
approach and speeds hypothesis development
Visual comparison of genome-scale datasets supports:
Faster and broader evaluation of the datasets
Identification of biases and assumptions in our methods
Novel insights into biological processes  new and more
focused questions
The Biological Process: the yeast cell cycle
VxInsight clustering of exit from stationary phase data:
T=0, 15, 30, 45, and 60 minutes after re-feeding
Ribosome ridge
Stationary-Phase genes