Biomedical Engineering Reference
In-Depth Information
8
Using microarrays to measure cellular
changes induced by biomaterials
V. M. ARIS, University of Medicine and
Dentistry of New Jersey, USA
DOI
: 10.1533/9780857093684.262
Abstract
: Microarrays represent the miniaturization and high throughput
serialization of a laboratory assay. There are several types of diverse array
technologies currently in use to measure changes in gene expression.
In this chapter, we will underscore how to design experiments in order
to be able to answer questions posed by the investigator, the different
normalization strategies used and their assumptions, the different types
of statistical analysis and adjustments for multiple testing, methods of
uncovering patterns in the data, and strategies to fi nd meaning in the
results.
Key words
: microarrays, gene expression, experimental design,
normalization, statistical analysis.
8.1 Introduction
Microarrays, sometimes called multiplex lab on a chip, were developed to
miniaturize laboratory assays increasing the throughput and reducing cost
and needed material per assay. There are many types or arrays perform-
ing many different tests. One can array on a solid substrate DNA oligo-
nucleotide, proteins, peptides, tissues, cells, chemicals, antibodies, etc. The
tests performed will measure very different biological/chemical responses.
Regardless of the array nature, there are some common key points the exper-
imenter will need to address in order to design and analyse his experiment.
In this chapter, we will discuss the different experimental design and what
question they can and cannot answer, the different normalization strategies
used and their assumptions, the different types of statistical analysis and
adjustments for multiple testing, the different pattern recognition methods,
and the strategies to fi nd meaning in the results.
Microarray assays are usually arranged on a two-dimensional solid sub-
strate such as a glass slide or a thin-fi lm silicon cell for the purpose of test-
ing small amounts of biological material. Data analysis depends greatly on
262
