Biology Reference
In-Depth Information
(a)
T
2
T
1
T
3
Microarrays
T
4
T
5
Individual
Experiment
'Small'
dataset
(b)
Cell
cycle
Gene
deletion
Stress
KO
+
+
+…=
Pooling of profiles
Large
heterogeneous
dataset
Fig. 1.
Expression data (a) Individual microarray experiments addressing specific
biological issues give rise to small datasets comprising only a few distinct experi-
mental conditions (e.g. time points). (b) Large-scale expression data can be gener-
ated by pooling profiles from many such individual experiments (or by conducting
dedicated comprehensive assays). Such data cover not only thousands of genes, but
also many cellular states by including a heterogeneous collection of experimental
conditions.
expression, and dimension-reduction techniques (reviewed by Brazma
et al
.
4
and Slonim
5
).
In addition to the specific biological questions addressed in such
individual focused experiments, it is widely recognized that a wealth of
supplementary information can be retrieved from a large and heteroge-
neous dataset describing the transcriptional response to a variety of dif-
ferent conditions.
2
Furthermore, the relatively high level of noise in these
data sets can be dealt with most effectively by combining many arrays
