Information Technology Reference
In-Depth Information
cDNA micro arrays are capable of profiling gene expression patterns of tens of
thousands of genes in a single experiment. DNA targets are arrayed onto glass slides
or membranes and probed with fluorescent- or radioactively-labelled cDNA [5].
Interesting is the image analysis and data extraction. The highly regular arrangement
of detector elements and crisply delineated signals that result from robotic printing
and confocal imaging of detected arrays renders image data amenable to extraction by
highly developed, digital image processing procedures. Mathematical morphology
methods can be used to predict the likely shape and placement of the hybridization
signal. In contrast, extraction of data from film or phosphor-image representations of
radioactive hybridisations presents many difficulties for image analysis.
2.2 Data Processing Phase
Data processing starts with the phase of digitalizing (Fig. 2). A scanner is used for
digitalizing arrays. Digital data are stored in a database - a central part of the whole
laboratory information management system (LIMS) which manages hybridization
results in general. All array methods require the construction of databases within a
LIMS for the management of information on the genes represented on the array, the
primary results of hybridization and the construction of algorithms to examine the
outputs of single or multiple array experiments.
Data analysis in this phase is often based on correlation methods developed for the
analysis of data which is more highly constrained than at the transcript level.
Scanning and image processing are currently resource-intensive tasks, requiring to
ensure that grids are properly aligned. Artefacts have to be labelled and properly
excluded from analysis. Standard input and output formats have to be fixed,
automation of identifying features and artefacts have to be made sure. Providing
routine quality assessment and the assignment of robust confidence statistics of gene
expression data is not simple. The quality assurance information should be
transmitted with the primary data . The basic idea is to reduce an image of spots of
varying intensities into a table with a measure of the intensity for each spot. There is
as yet no common manner of extracting this information, and many scientists are still
writing customized software for this purpose. A variety of software tools have been
developed for use in analysing micro array data and processing array images [6] .
Some tools for analysing micro array data are available at http://genome-
www5.stanford.edu/MicroArray/SMD/restech.html . This software is provided by
micro arraying groups from Stanford. Especially a tool for processing fluorescent
images of micro arrays called ScanAlyze is provided for free download.
Technologies for whole-genome RNA expression studies are becoming increasingly
reliable and accessible. Universal standards to make the data more suitable for
comparative analysis and for inter-operability with other information resources have
been yet to emerge [2]. In carrying out comparisons of expression data using
measurements from a single array or multiple arrays, the question of normalizing data
arises. There are essentially two strategies that can be followed in carrying out the
normalization: to consider all of the genes in the sample or to designate a subset
expected to be unchanging over most circumstances. In either case, variance of the
normalization set can be used to generate estimates of expected variance, leading to
predict confidence intervals. A lot of explicit methods have been developed which
make use of a subset of genes for normalization, and extract from the variance of this
Search WWH ::




Custom Search