John M. S. Bartlett.pdf - Bio-Nica.info

Differential Display Techniques 221
can be examined simultaneously in a single experiment and the rapidity with which
information can be gathered from multiple samples.
2.3. SAGE
In SAGE, short expressed sequence tags are produced from each mRNA expressed.
Thus far, the technique is very similar to that of expressed sequence tag (EST) library
screening. The advantage of SAGE is that the tags used are 9- to 10-bp long, and
through a number of steps (see chapter 40 by Oien, these tags are concatenated into
clones containing multiple (up to 40+) sequence tags before sequencing. Therefore, the
expression profile is generated by sequencing many such clones and simple counting
of the number of times each tag is represented (26).
Therefore, the chief difference between DNA microarrays, differential display, and
SAGE is that SAGE produces a digital count of the number of clones representing
each sequence expressed. This is, however, only one of the differences between these
approaches. Microarrays can identify large numbers (10,000 or more) of expressed
genes, and determine, semiquantitatively, alterations in their expression. However,
genes whose expression is modulated only marginally are unlikely to be identified by
this approach. In SAGE, each expressed gene may be sequenced and thus single copy
changes in expression may be detected if sufficient clones are sequenced. In SAGE,
the sensitivity is determined not by the detection system but by the amount of effort
given to identification of sequence tags. A further advantage of SAGE analysis over
microarrays is that it is not necessary to have to hand a clone representing the gene(s) of
interest; because the detection of genes is performed by sequencing, no hybridization
matrix is required. SAGE analysis is therefore limited only by the amount of sequencing
that can be performed in a cost-effective manner. Conventional sequencing of ESTs
relies on the analysis of several hundred bps to identify each gene. In SAGE the
sequence tag used to identify the expressed sequence is 10 bp. Theoretically, the
discriminatory power of a 10-bp region of cDNA is 1 in 1,048,576 (4 10 ). By linking
together the sequence tags from many different genes into a concatemeric clone, a
single sequencing run of 800 bp can be used to identify 50 or more different sequences.
Given the high throughput available in today’s 96-lane sequencing platforms, approx
5000 gene transcripts may be analyzed from a SAGE library in a single sequencing
run. In addition, SAGE has the advantage that digital data is more suited to statistical
analysis. Finally, and perhaps of greatest significance as increasingly experiments
become more time consuming and costly to perform the quantitative nature of SAGE
library analysis potentially allows direct comparisons between laboratories and between
libraries, subject only to the constraints of the initial experimental variables. Initial data
comparisons support the premise that comparisons between laboratories performing
SAGE can provide valuable and valid information. The caveat that must be applied
is that although such comparisons are at present validated by internal controls and
performed in expert laboratories with rigorous controls, there are also specific pitfalls in
SAGE analysis relating to identification of clones and PCR biases in construction of tags
(26). Although some groups are already extrapolating from this data to compare data
from different experimental approaches (27), in the light of some concerns raised and
the possible impact on gene expression of even small changes in pH, nutrient, etc., this
approach is at present premature, not just for SAGE but for all expression analyses.