Jim Tokuhisa wrote:
>> Firstly, I want to thank Brian for the time and effort spent
> to arrange the 'genearray' group. Now it is up to the rest
> of us to make this a useful resource.
>> Our group is planning to buy an imaging system that will be
> used in part for the analysis of macroarrays.
>> What imagers have folks used and how satisfying have the
> results been? With the results, have there been successes
> or difficulties that you would attribute to particular
> features of an imager? Any comments on repair histories,
> the ease of use of the hardware or especially the software
> would be appreciated.
>
Rather than advocating any specific imager, I want to
put in my 2cents worth on what I consider to be an important
part of the strategy. That is, don't worry too much about
having integrated software that does everything from
image acquisition to managing the database to mining the
data. Here are the reasons:
1) Image acquisition is a very short term process that is done
(hopefully only onece), while data management and analysis are
long term projects, that in some senses are never done. If
everything is done on a single PC, from start to finish, then
you have a line of people waiting to analyze their data
while somebody else is reading slides.
2) Data warehousing and analysis should be server-based.
Software should reside on the server, so that all
workers can access it. Then the only consideration is
how many licenses to buy.
3) Separation of the data acquisition task from the
analysis task frees you up to buy the best hardware,
rather than the best hardware/software combination.
The hardware is a very expensive long-term investment.
The software is cheaper, so you have the luxury of
maybe even buying 2 or 3 programs to try out, before
settling on one. With software, it's also important
to note that no one package will ever do everything,
and no one package will do all tasks that it does
do well.
At the same time, you want to be free to acquire
new software, both commercial and non-commercial,
When new analytical methods are developed.
The less your strategy is wedded to a particular
piece of hardware, the better.
--
=============================================================================
Brian Fristensky | In Paradise there are no stories,
Department of Plant Science | because there are no journeys. It's
University of Manitoba | loss and regret and misery and
Winnipeg, MB R3T 2N2 CANADA | yearning that drive the story forward,
frist at cc.umanitoba.ca | along its twisted road.
Office phone: 204-474-6085 |
FAX: 204-474-7528 |
http://home.cc.umanitoba.ca/~frist/ - Margaret Atwood, THE BLIND
ASSASSIN
=============================================================================
