Best way to study genome-wide gene expression?

Austin So (Hae-Jin) haejin at netinfo.ubc.caX
Sun Dec 10 19:49:20 EST 2000

It really depends on two things: how much money you have at your disposal
and what kind (quality and quantity) of information you are interested in.

Both microarray and SAGE actually require a lot of genomic sequence
information to be known in order to be used to their fullest
potential...though you can get a lot of information from both.

You can't identify SAGE tags unless you have quite a bit of sequence
information, and the biggest drawback to SAGE is the depth in sequence
required in order to begin to pick out low copy messages and the cost or
your sequencing runs. Although this in itself is not too bad, the problems
occur when there isn't a unique tag for every gene, which you can't sort out
if you have no sequence information. But you could just simply go back to
your clone library, and pick out those clones that exhibit a change using
your SAGE tag...but then, you could have just done a subtractive libary,

The big drawback to microarrays is the overhead cost (at minimum, arrayer
and reader), but you are working with cDNA clones which allow you to simply
go back to your clone and check things out. And the fact of the matter is
that microarray data should really only be considered qualitative.

But there is no greater requirement for sequence information between the
two, and in fact among a lot of "global" expression tools.

"Melanie Chang" <Melanie_Chang2000 at> wrote in message
news:iJPY5.7663$c32.1213543 at
> Hi,
> I have recently relocated to a new company in the Bay Area and my boss
> me to look into setting up a genome-wide gene expression profiling
> The problem - we are working on human parasites, and none of their genomes
> have been sequenced so far or will be in the near future. Gene
> chips/microarrays are therefore out of question for the time being.
> What other technologies are out there that I should take a closer look at?
> have heard of LEAD, SAGE and TOGA (as done by DGT)? How solid do they
> perform and are they licensable? Are there better ways?

Austin P. So (Hae Jin)

Biotechnology Laboratory
University of British Columbia

E-mail: haejin at (under construction)


More information about the Methods mailing list