Transcript Tmm - OpenWetWare

Tmm: Analysis of Multiple
Microarray Data Sets
Richard Moffitt
Georgia Institute of Technology
29 June, 2006
Goal
• Use 60 large human microarray datasets.
(3924 arrays)
• Find reliably coexpressed genes.
• http://benzer.ubic.ca/cgi-bin/find-links.cgi
– (just google ‘tmm microarray’)
Usage Case
• Query by
gene or
probe ID.
• Set
stringency
level.
How it Works
• Looks for genes that coexpress with the
queried-for gene.
– correlates gene expression profiles
• Stringency requirement eliminates weak
links.
Our Query
• RAP1GSD1, a biomarker form Chang et al
• 2 minutes later…
Our Results
• List of linked
genes and
some
statistics.
Visualization
• Visualizations of
coexpresed gene
profiles for each
dataset used.
Query #2
LETMD1, a biomarker from
Citation
Spira A, Am J Respir Cell Mol Biol. 2004
Phenotypes_Being_Studied
No or mild emphysema, severe emphysema
Chip_Platform
GPL96: Affymetrix GeneChip Human Genome U133 Array
Set HG-U133A for 712X712
Results #2 :
Why?
• Our first query was from one of the
datasets used by Tmm.
Synonym Search
Conclusion
• Useful to make a small list of probable
targets.
• Useful for some validation?
– Similar to GOMiner validation.
– Speed will inhibit this.
• Semantics is a barrier to usefulness.
Acknowledgements
• Thanks to:
Deepak Sambhara
JT Torrance
Lauren Smalls-Mantey
Malcolm Thomas
Randy Han
and Kiet Hyun
for curating all the biomarker data that was used
for test queries.