|
Terrence S. Furey
answers a few questions about this month's
new hot paper in the field of Computer Science.
From
•>>January 2003
Field: Computer Science
Article Title: "Support vector machine classification and validation of cancer tissue samples using microarray expression data"
Authors: Furey,
TS;Cristianini, N;Duffy, N;Bednarski,
DW;Schummer, M;Haussler, D
Journal: BIOINFORMATICS
Volume: 16
Page: 906-914
Year: OCT 2000
* Univ Calif Santa Cruz, Dept Comp Sci, Santa Cruz, CA 95064 USA.
* Univ Calif Santa Cruz, Dept Comp Sci, Santa Cruz, CA 95064 USA.
* Univ Bristol, Dept Engn Math, Bristol BS8 1TH, Avon, England.
* Univ Washington, Dept Mol Biotechnol, Seattle, WA 98195 USA.
|
 Why
do you think your paper is highly cited?
Diagnosing
cancer is obviously a subject of intense research, and any new
technology that aids in this endeavor is of great interest.
Advances in microarray technology have provided a new
large-scale method in which to analyze tissues at a molecular level,
and a high-powered classifier is needed to interpret this data and to
provide a computational diagnosis for a particular sample.
The classifier we report on, a support vector machine (SVM),
is a fairly new machine learning technique, and we show
that it can perform this task very well.
So, the combination of new biological technology with a recent
machine learning technique applied to a critical problem makes this
paper interesting in many respect s.
Does
it describe a new discovery or a new methodology that's useful to
others?
The
paper describes a novel method of classifying tissues based on
microarray expression data.
We show that the SVM
classifier performs well using multiple data sets including a novel
dataset for ovarian tissues.
We also show that SVMs can be used to detect mis-classified samples.
Given the subject matter, we believe this is of interest to
many researchers, and that our general methodology will work on many
types of microarray datasets.
What
were some of the circumstances that led you to do this research?
We
had been closely following the developments in microarray technology,
and had previously analyzed yeast datasets using SVMs.
Richard Karp was instrumental in putting us in touch with
Michel Schummer, who was using microarrays to look at ovarian tissues.
With the additional public availability of datasets on colon
tumors and leukemia, it was natural to apply this recent machine
learning algorithm to analyze these data.
Could
you summarize the significance of your paper in layman's terms?
Currently,
the determination of whether a particular tumor is benign or malignant
is fairly subjective.
Cancer also comes in many forms, even within the same type of
tissue, and thus response to different treatments will vary for each
case. The
information from these microarray experiments allow a much more
detailed look at what is going on in tumor cells.
With this data and special computer applications, a more
objective and accurate diagnosis may be achievable.
In addition, it may be possible to characterize more types of
cancer and be able to better predict how each will respond to
different treatments.
Terrence
S. Furey
Postdoctoral Researcher
Howard Hughes Medical Institute
Center for Biomolecular Science and Engineering
University of California, Santa Cruz
Santa Cruz, CA, USA
|
ESI Special Topics,
January 2003
Citing URL - http://www.esi-topics.com/nhp/2003/january-03-TerrenceFurey.html
|
|
