thanks Rafael,
another question, I guess GCRMA is going to be memory consuming in a
more aggressive way than RMA ? is there a GCRMAexpress planned ?
I m currently testing it and it is for sure less fast than RMA... with
8 chips (HU133A) it takes more time than with 40 chips under RMA ! By
the way I have compared 2 differential analysis using either RMA or
GCRMA all the other steps being done the same way with the same chips
and some genes had some dramatic changes with GCRMA... SInce it will
not be possible to check all these by PCR, is there any way to say
which method is the most reliable one and reflects "reality" ? I bet
you are going to say GCRMA...! Biological knowledge about these genes
is the only way for me to go ahead but...it is limited !
Thanks
Philippe
we are constantly improving gcrma and until we have a final product,
there will be no fast version.
prelimary results with spike in experiments suggest that gcrma is a
bit
more accurate than RMA. i feel quite optimistic about the model its
based
on, but the software implementation has not been tested thoroughly on
other data sets. criticism/comments/suggestsion are greatly
appreciated.
for this particular package you can interpret the version number
smaller than 1.0 as an indication that it is still under development.
On Wed, 27 Aug 2003 Phguardiol@aol.com wrote:
> thanks Rafael,
> another question, I guess GCRMA is going to be memory consuming in a
more aggressive way than RMA ? is there a GCRMAexpress planned ?
> I m currently testing it and it is for sure less fast than RMA...
with 8
chips (HU133A) it takes more time than with 40 chips under RMA ! By
the
way I have compared 2 differential analysis using either RMA or GCRMA
all
the other steps being done the same way with the same chips and some
genes
had some dramatic changes with GCRMA... SInce it will not be possible
to
check all these by PCR, is there any way to say which method is the
most
reliable one and reflects "reality" ? I bet you are going to say
GCRMA...!
Biological knowledge about these genes is the only way for me to go
ahead
but...it is limited !
> Thanks
> Philippe
>
