Rosetta Resolver News

You are not logged in [Log in] [Register]

Questions & Answers [change]

Log(error)
Ellen Sterrenburg <E.sterrenburg@lumc.nl>
2004-04-15 17:57:37
Hi Wayne,
I'm using the log(error) to see if the specific spot on the array is
expressed.
Everything higher than 0.1 is regarded as not expressed.
But the lowest log(error) that I have is 0.1.
Is this normal, or can it be lower?
Best regards,
Ellen

Hi Ellen,
Log(error) in a ratio profile is the estimated standard error of
log10(ratio)=log10(RedIntensity/GreenIntensity). It is the log-ratio error.
It is in log10 scale. The GenePix tech note describes how we compute the
log-ratio error in the feature level. Then features for the same reporter
will be squeezed (averaged) together in reporter-level profiles. In the
squeeze process there is a hard limit to the lowest level of log-ratio
error, 0.1. Anything below the level will be set to the level. Even if there
is only one feature for a reporter, it still goes through the same squeezing
process. The low limit still applies.

Log-ratio error tells us the uncertainty range of the log-ratio measurement
for a given microarray technology. But we definitely should not use the
log-ratio error alone to define whether a gene is differentially expressed
or not. The ratio p-value takes both log-ratio and log-ratio error into
consideration. We should use p-value to make the differential call.

Ellen:
So, in conclusion the log(error) does NOT say anything about if a transcript is expressed or not. It just says something about the spread in the data of that transcript. And for not expressed transcripts, this spread is often bigger (more inconsistent), and that is why you get a higher log(error).
[reply to this message]

[new message]
Title
Author
Date
Heat 3287: flags in GenePix -> fail codes in Rosetta
Judith Boer
2003-03-03 12:04:44
Heat 3358: export data for several profiles/experiments ...
Judith Boer
2003-03-03 12:05:19
Heat 3288: import_gpr-profiles.txt
Judith Boer
2003-03-11 11:29:45
export-anova.txt
Judith Boer
2003-03-11 11:30:13
adding columns of gene annotation to expression data
Judith Boer
2003-03-11 11:31:33
only ~4000 UniGene clusters Affy chip U133A?
Judith Boer
2003-03-11 11:34:10
Heat 3573: Changing the scale of y-axis in trend plots
Ellen Sterrenburg
2003-03-12 16:47:55
problems with image view
Peter-Bram 't Hoen
2003-03-14 14:44:20
Re: problems with image view
Peter-Bram 't Hoen
2003-03-28 11:07:43
Re: problems with image view
Peter-Bram 't Hoen
2003-03-28 11:46:50
Re: problems with image view
Ellen Sterrenburg
2003-04-03 17:56:18
dealing with flags
Peter-Bram 't Hoen
2003-03-14 15:33:43
do not use dashes in chip names!
Peter-Bram 't Hoen
2003-03-28 11:09:53
importing gpr files
Peter-Bram 't Hoen
2003-03-28 11:44:33
Re: importing gpr files
Ellen Sterrenburg
2003-04-03 17:54:08
Dye-swaps and combining hybs
Ellen Sterrenburg
2003-04-04 16:09:04
Importing a whole list of Acc. # to create a bioset
Ellen Sterrenburg
2003-04-09 14:57:22
Use of self-self hybridization in an ANOVA
Ellen Sterrenburg
2003-04-09 14:59:47
Heat 3903: "stitch" together data of two chips ...
Ellen Sterrenburg
2003-04-10 11:46:34
Heat 3954: Determine the number of clusters for K-means ...
Ellen Sterrenburg
2003-04-15 17:35:41
Heat 3902: Clustering in trend viewer
Ellen Sterrenburg
2003-04-15 17:39:28
Heat 3963: 1 table with cluster number and ratio data
Ellen Sterrenburg
2003-04-17 07:44:13
Heat #3965: Add extra colum with GO info
Ellen Sterrenburg
2003-04-23 07:49:51
Heat 4053:Select genes with signal in at least one exp. ...
Ellen Sterrenburg
2003-04-30 07:47:48
Heat 4211: Request for colored clusters from trend viewer
Ellen Sterrenburg
2003-06-17 07:42:37
Heat 4789:Spots with no signal
Ellen Sterrenburg
2003-07-11 11:21:19
Re: Heat 4789:Spots with no signal
Ellen Sterrenburg
2003-07-14 15:04:57
What to do before you load your profile?
Ellen Sterrenburg
2003-08-18 15:36:28
cell number
Peter-Bram 't Hoen
2003-08-27 17:07:29
image viewer - zoom in?
Peter-Bram 't Hoen
2003-08-27 17:10:25
polarity
Peter-Bram 't Hoen
2003-08-27 17:11:05
Correcting for multiple testing in Rosetta
Ellen Sterrenburg
2003-09-25 17:59:08
Q$A: How do you delete data from the system?
Anders
2003-10-22 10:11:08
Import Agilent pattern
Peter-Bram 't Hoen
2003-11-13 14:53:40
downregulated features that appear as upregulated
Peter-Bram 't Hoen
2003-11-13 14:55:50
Export data in Excel format
Ellen Sterrenburg
2004-01-27 12:27:06
Re: Export data in Excel format
Ellen Sterrenburg
2004-02-11 07:43:51
Plotting and exporting ANOVA results
Ellen Sterrenburg
2004-02-11 07:53:03
The export-values in the trend are always log10 values!!!! ...
Ellen Sterrenburg
2004-02-19 07:52:59
Log(error)
Ellen Sterrenburg
2004-04-15 17:57:37
Exporting high-quality figures for publications
Peter-Bram 't Hoen
2004-08-27 13:07:47
ANOVA in experiment defenition-multiple testing
Peter-Bram 't Hoen
2005-02-09 09:03:41
Combine-on-the-fly of experiments no longer possible
Peter-Bram 't Hoen
2005-04-13 10:21:54
Hainofaft

2009-01-13 21:07:52
Re: Hainofaft
Ivo Fokkema
2011-01-17 11:30:46