Tissue Specificity

[water...@peds.ucsf.edu]

Hello,

My question is regarding the output of tissue specificity for a given gene on BioGPS.  My understanding from previous posts (see below) is that depending on the probe used for a specific gene the output can vary and that it is difficult to compare multiple probes for a given gene because they may give different tissue-specific patterns of gene expression.  Thus, I would like to know how biogps determines which probe to use to come up with the tissue-specific tissue destribution for a given gene on the website display?  Is the Biogps readout just one probe or is it the synthesis of multiple probes for a given gene?  How does Biogps determine which probe to display and is there any system used to determine the best probe for a given gene based on the microarray data?  Thanks in advance for your help.

 

Mike

 

Comparing expression of genes within a single tissue Another discussion addresses this topic: You were redirected here from a topic that was marked as a duplicate: 2 posts by 2 authors Previous Previous Next Next

Andrew Walsh

11/17/09

Other recipients:

Hello,

I was wondering if there is any way to compare expression of genes within a single tissue. From browsing the group messages, it seems that this is not possible. I had assumed it was, but I think my confusion may be related to having thought that this paper was relevant to the expression data being shown:

http://www.pnas.org/content/99/7/4465.abstract?sid=f85a9054-89bb-4466-a3eb-fa5721d1b330

"""
Identification of Tissue-Specific Genes. For the human dataset, the
set of 46 tissues, organs, and cell-lines was reduced to 25
independent and nonredundant samples (see Table 1, which is
published as supporting information on the PNAS web site,
www.pnas.org). All 45 mouse samples were derived from dis-
section and were already considered as having independent
origins. Based on extensive PCR-validation of oligonucleotide
array data (data not shown) and the absence presence call
provided by the GENECHIP software package, an average differ-
ence (AD) value of 200 was defined as a conservative threshold
to call a gene ‘‘expressed’’ or present. Additionally, an AD of 200
has been estimated to represent 3–5 copies per cell, and an
expression ratio of 2-fold has previously been established as the
approximate limit of sensitivity (10, 11). By using these guide-
lines as filtering criteria, tissue-specific genes were conserva-
tively defined as having an AD value of greater than 200 in one
tissue, and AD value of less than 100 in all other tissues.
"""

Apparently this does not apply to the Symatlas data (at least as far as I could gather reading through the board, though perhaps I am wrong). I can get by without having specific copy numbers, but it would be helpful to get even approximate rankings. I've noticed that the probe sets for one gene can have very different values. For example, for LDLR in HeLa cells, the intensity values for the probe sets are very different:

Tissue

Tissue_label

217005_at

217183_at

217173_s_at

202067_s_at

217103_at

202068_s_at

HELA

0028_HELA

6.7

3.56

10.4

33.13

7.21

303.81

Even if the intensity values are highly variable and I cannot come up with a copy number, is there some some way that I can use the Symatlas data to produce an approximate ranking of gene expression within tissues?

Thank you very much for your time and help,

Andrew

---

Get a great deal on Windows 7 and see how it works the way you want. Check out the offers on Windows 7now.

Show trimmed content

Click here to Reply

Andrew Su

11/17/09

Other recipients: Andrew Walsh...@hotmail.com>

Hi Andrew,

Because of the way that the Affymetrix chips work, comparisons of microarray data are best made for a single probe set across many tissues. The type of comparison you're talking about (across many probe sets in a single tissue) is significantly less reliable. As you note below, different probe sets for the same gene can often have very different distributions, likely due to different background signals, cross-hybridization, secondary structure, etc. You are correct that this is exactly the type of analysis we talk about in the quoted passage below, but we purposely only make very modest interpretations of those results.

Hope that helps...

Cheers,

-andrew

[Andrew Su]

Hi Mike,

When multiple probe sets match to a single gene, BioGPS shows them all.  You can switch between them using the Probeset drop down box (see attachment).  As for which probe set is shown by default, BioGPS does the most naive thing possible and shows the first one alphabetically.  

Hope that helps...

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
BioGPS: http://biogps.org/
Blog: http://sulab.org/blog/
Twitter: http://twitter.com/biogps
 
---
You received this message because you are subscribed to the Google Groups "BioGPS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to biogps+un...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.