245

Chapter 6

Gene Expression Pattern Discovery

fferentially expressed genes are treated as the major signal

understand how a stress results in a phenomenon change in

species. Scientists hypothesised that it is the genetic

nature that ultimately causes the observed phenomenon

ange. The major objective of gene expression pattern

covery is thus to discover differentially expressed genes.

this chapter, the approaches for gene expression pattern

covery will be introduced for two major types of data, i.e.,

microarray gene expression data and the RNA-seq

quencing count data. Gene expression pattern discovery

y encounter the challenges of the outlier problem or the

bpopulation problem. Therefore this chapter will also

roduce the alternatives for gene expression pattern

covery. Moreover, insufficient replicate number is the

mmon issue in gene expression pattern discovery. The

al-scale Gaussian mixture algorithm for discovering

ferentially expressed genes in a data set with insufficient

licates will be introduced in this chapter.

rentially expressed genes

situations, an experiment with two conditions is conducted to

which has caused a phenomenon change in a species for it to

to a stress. These two experimental conditions normally have

settings. One is called the control condition, in which no stress is

nd the other is called the case condition, in which a stress is