Biomedical Engineering Reference
In-Depth Information
stance, if we are seeking a set of diagnostic markers we should probably look for
a small yet robust set of genes that strongly differentiates between case and con-
trol. This is the basis for DNA array-based diagnostics, of which we gave a suc-
cessful example in the previous section. In this context, if we keep adding
markers, the performance of the classification will eventually deteriorate. How-
ever, if we are interested in discovering biological processes that behave differ-
entially between case and control populations, we should probably be rather
encompassing in our differentiating features.
The lists of genes generated by gene selection methods will have to be or-
ganized with the help of literature search techniques, ontologies, or other
sources of biological information. These lists will likely trigger hypothesis-
driven research, which will give rise to a more mechanistic understanding of
basic biological processes and of the ways in which these cellular processes are
affected in disease.
6.
ACKNOWLEDGMENTS
I am deeply indebted to my collaborators Yuhai Tu, Andrea Califano, Jorge
Lepre, and Ulf Klein for many discussions on how to best analyze gene expres-
sion data. Thanks also go to Riccardo Dalla-Favera for providing the CLL,
DLBCL and FL data from Columbia University, and Diane Jelinek and Neil
Kay for providing the Mayo Clinic CLL data, and for their support on the bio-
logical side of the work presented in this chapter.
7.
NOTE
1. For nonprofit or academic institutions, a free version of Genes@Work
can be downloaded for noncommercial research purposes only from
http://www.research.ibm.com/FunGen.
8.
REFERENCES
1. Stolovitzky G. 2003. Gene selection in microarray data: the elephant, the blind men and our
algorithms.
Curr Opin Struct Biol
13
:370-376.
2. ArrayExpress database on World Wide Web: http://www.ebi.ac.uk/arrayexpress/
3. Stanford Microarray database on World Wide Web: http://genome. www4. stanford.edu/
MicroArray/SMD/
4.
GenomeWeb Gene Expression and Microarrays on World Wide Web: http://www.hgmp.mrc.
ac.uk/GenomeWeb/nuc-genexp.html
5.
YF Leung's Microarray Links on World Wide Web: http://ihome.cuhk.edu.hk/%7Eb400559/
array.html
