Download Normalized Data
The primary data sets contains the expression levels (average
difference between perfect match and mismatch oligos) and a accuracy call
(P, M or A; more details at
Affymetrix and
Wodicka et al., 1997). Four expression profiles are
available for each transcription factor studied: two wild type and
two mutant. Each mutant profile has been normalized to each wild
type profile by a scaling factor based on the levels returned for 5
different control RNAs that were spiked into the total RNA
preparations. Each control RNA was represented by 3 different sets of
oligos on each of the 4 chips that make up a single genome. The first
column contains the gene name used by Affymetrix. The second column
contains the ORF name or is blank if no standard ORF name could be
produced from the Affymetrix gene designation. This is followed by
the expression level and accuracy call for each of the four profiles:
column 3 is the first wild type expression level, column 4 is the first wild
type accuracy call, column 5 is the first mutant expression level, column
6 is the first mutant accuracy call, column 7 is the second wild type expression
level, column 8 is the second wild type accuracy call, column 9 is the second
mutant expression level, and column 10 is the second mutant accuracy call.
These values are separated by tabs, with a carriage return following
the last value for a given gene.
It should be noted that the confidence calls reported in the
website reflects the confidence in the reported fold change. This is
related to, but does not necessarily represent, the confidence in
whether that particular gene is directly regulated by the
transcription factor studied. For example, the criteria for an
"excellent" confidence in the fold change was that a "P" call was
returned for an mRNA level in all four profiles. If an mRNA level
dropped below the accurate level of detection in the mutant profiles,
this would result in two "P" calls for the wild type levels and two
"A" calls for the mutant levels. In this case the confidence in the
reported fold change is not "excellent" but "good", even though that
gene may be an excellent candidate for being directly regulated by
the transcription factor because a similar drop was observed in the
experiment performed with the temperature-sensitive mutation in core
RNA polymerase II. In such cases a conservative estimate of the
actual fold change can be made by flooring the level to an arbitrary
"noise" level of 10 - 20 units.
Download normalized data for FCP1
(Save the window's contents as 'Source', not as 'Text', for importing into your favorite program.)