2 questions about bioGPS
Big Boss
Thank you for running this web site.
My first question is about the expression data, which are represented on-line graphically with error bars, but exist as single values in the downloads. What do the error bars represent, and could the relevant information be included in the gene-specific download? My concern is about many cases when the highest expressing sample also has very wide error bars, which (depending on what they represent) cold make it statistically indistinguishable from the low-expressing ones. Thus, presenting a single value (average?) would give a completely wrong impression.
My second question is about "Reporters". I am trying to figure out why there are five reporters for the same gene (human PAK2) on the same array (U133A) and why, at least, some of them show very different picture. I can only guess that this represnts alternative transcripts. Would you be able to provide a direct link to a page where entering the reporter's name (as shown on your site) would give more information about it? Perhaps, it is possible to have the reporter name as a "clickable" item with a link to additional information?
Thank you!
Eugene.
Chat with Messenger straight from your Hotmail inbox. Check it out
Andrew Su
Hi Eugene,
You offer two excellent suggestions below. I've logged them both in our feature request system. Both are very feasible, but I don't yet have a specific timeline for when they'll be rolled out.
On the second issue, a work-around that you might consider now would be to create a custom layout containing the UCSC genome browser plugin. UCSC has all the Affymetrix reporters mapped to the genome assembly. See the screenshot below on how you can do this, and feel free to ask if you have any problems with this.
For being the first to suggest these features, can I offer you one of our BioGPS mascots? (http://biogps.blogspot.com/2008/12/biogps-swag-just-in-time-for-holidays.html). Please send me your selection and mailing address off-list.
Cheers,
-andrew
--
You received this message because you are subscribed to the
Google Groups "BioGPS" group.
To post to this group, send email to bio...@googlegroups.com.
To unsubscribe from this group, send email to
biogps+un...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/biogps?hl=en.
Andrew Su
Hi Eugene,
I should also have mentioned that another way to access the raw data is to download the entire data matrix from our Downloads page (http://biogps.gnf.org/downloads). These data files do not have the replicates averaged.
Cheers,
-andrew
Eugene
I followed your link and got the Human U133A/GNF1H Gene Atlas gnf1h-
gcrma.zip file. However, it looks like it has a single value per
reporter. Did I get a wrong file?
Also, if I understand it correctly, the variability displayed by the
error bars comes not from multiple replicates, but from non-identical
behavior of separate probes or probesets. Is this right?
Eugene.
Andrew Su
Apologies, I thought we'd posted the unaveraged file, but you are correct that the currently linked file has averaged values. We'll post a link to the unaveraged file on the downloads page soon, but in the mean time, you can download it using this direct link:
http://plugins.gnf.org/download/gnf1h-gcrma-unaveraged.zip
The replicates for this data set were prepared as follows:
"When samples from four or more subjects were available, equal numbers of male and female subjects were used to make two independent pools; when fewer than four samples were available, RNA samples were pooled, and duplicate amplifications were performed for each pool (detailed annotation for human samples is on our web site and in Table 1, which is published as supporting information on the PNAS web site)."
Additional details can be found in the PNAS 2004 paper listed on the downloads page. To emphasize however, we never average values obtained from separate probe sets.
Hope that helps.
Cheers,
-andrew
Andrew Su
Hi Eugene,
... and I wanted to let you know that we just released a minor update to BioGPS. The Affymetrix reporters in the "Gene Identifiers" plugin are now directly hyperlinked to UCSC.
Cheers,
-andrew
From: Big Boss
[mailto:myow...@hotmail.com]
Sent: Sunday, December 06, 2009 8:59 PM
To: bio...@googlegroups.com
Subject: [BioGPS] 2 questions about bioGPS
Hi!
--