The Cancer Genome Anatomy Project

[swimmy]

I tried the experiment called "The Cancer Genome Anatomy Project" but I don't get what you were supposed to find out. It seems like a demonstrative project, not an experimental one. Is it meant to be experimental? If it isn't does anyone have any ideas of how it can be an experimental project?

Thanks,
swimmy

[MelissaB]

Hi,

This is indeed a demonstrative project. In order to make it into an experimental one, I suggest coming up with an hypothesis--for example, 'Gene X will be widely expressed in many types of cancer tissue' and then using the tools you learned to test your hypothesis. In order to do this, you might want to do some research on cancer cells, what they require, and what genes may provide what they require.

[swimmy]

Thanks for the help. I have another question about this project.
What does the link labelled "A<B" do?
(this is after the queries are submitted and the results are shown)
http://cgap.nci.nih.gov/SAGE/GetUserRes ... 1248046065
It seems like it is the list of overexpressed genes, but what is it exactly?

Also what if A is equal to B in the library but A is less than B in the tags?

Do the genes have to be only in Pool B to be overexpressed?

Thanks,
swimmy

[deleted-2574]

Hi swimmy,

i posed your question to NIH, and received the following response:

Thank you for contacting NCICB Application support! Regarding your question, “A<B” is a hyperlink to a list of genes whose expression in Pool A is less than their expression in Pool B. Conversely, “A>B” is a hyperlink to a list of genes who expression in Pool A is greater than their expression in Pool B.

Does the above answer your question? Please let me know. If it doesn't answer your question, I now have an NIH contact to ask additional questions.

[swimmy]

Thanks. I have another question. I don't really get the part that says "Figure 3. CLASP2 is over-expressed in pancreatic cancer. Blue indicates average expression level, red indicates high expression level. (CGAP website, n.d.)"
Does this mean that even if B>A it isn't automatically an overexpressed gene? So does it need to be red in the cancer section in the table for pancreatic cancer AND B>A?


Thanks,
swimmy

[MelissaB]

Right--just because B>A does not mean that statistically speaking, the gene is overexpressed. Generally, if the sample size is very low or the difference is not very big (say, 1 in A and 2 in B), we cannot be certain that that difference is not just due to chance. As an example, on average, men are taller than women. But if we just went out and measured the heights of one man and two women, we might actually conclude that women are taller than men if we measure the right people! So, you need to have a red gene and B>A in order to conclude that the gene is overexpressed.

[swimmy]

Thanks.
How about if the normal tissue is also red? Does that mean that the gene is not overexpressed?
Does the normal tissue need to be blue?
What does it mean if its pink or light blue? It says something about tags. I don't really get what that means.
Also, what can be figured out from overexpressed genes?
For example, one cure for cancer. I get what overexpressed genes are.
Can something be achieved in finding genes that are common to two types (chosen from Carcinoma, Sarcoma, Lymphoma and leukemia, Germ cell tumor, Blastic tumor or blastoma OR the different kinds of tissue) of cancer?
Hold on... Can the overexpressed genes be "turned off"?

Thanks,
swimmy

[MelissaB]

If the normal tissue is also red, then the gene is overexpressed in that tissue regardless of whether or not it is cancerous--so there is effectively no change, and that is not a gene you could use to test for cancer. The normal tissue does not need to be blue, but there needs to be a different color between the normal tissue and the cancerous tissue. The stronger the color difference (e.g. blue to red is a bigger difference than blue to pink), the more likely that gene is to have something to do with cancer.

Pink just means slightly overexpressed, while light blue means slightly underexpressed, probably.

In some cases, stopping expression of these genes would indeed stop the cancer, because cancer 'needs' these genes to be overexpressed. In other cases, it wouldn't matter. It is not as easy as it sounds, however! I suggest you do some research on gene regulation.

[deleted-2574]

Hi swimmy,

Below is a response I received from NIH to your questions. Numbers 8-11 require more work.

Cheers!

Dave
---------------
You might find useful information here:
http://cgap.nci.nih.gov/SAGE/SDGEDHowTo
 
Please find below some information and answers to your questions:

express SDGED:

Libs A (or B) The number of libraries which contain this tag
Tags A (or B) Tag frequency in either Pool A or B.
Tag Odds (A:B) The odds ratio uses a simple mathematical formula to provide a measure of the relative amount of a tag in pool A to Pool B.

The tag frequency in Pool A(or B) indicates the tag, or the gene, which mapped this tag, expression level in Pool A (or B)

A > B means that the tag frequency in Pool A which contain this tag is greater than the tag frequency in Pool B which contain this tag, i.e. the tag, or the gene mapping this tag over-expressed in A than B. We are talking two libs comparing.

1.  What does the link labelled "A<B" do?
(this is after the queries are submitted and the results are shown).
"A<B" ###ımeans that the tag frequency in Pool B which contain this tag is greater than the tag frequency in Pool A which contain this tag, i.e. the tag, or the gene mapping this tag over-expressed in B than A.

2.  Also what if A is equal to B in the library, but A is less than B in the tags?
See 1.: The gene mapping this tag over-expressed in B than A. We don’t care of the number of libraries of the Pool which contain this tag, whether they are equal or not, or reverse (i.e. the lib number of Pool A greater than lib number of Pool B, when A<B), when we discuss A>B or A<B.

3.  Do the genes have to be only in Pool B to be over-expressed?
No, if A>B the tag, or the gene mapping this tag over-expressed in A than B.

4.  In Figure 3, CLASP2 is over-expressed in pancreatic cancer. Blue indicates average expression level, red indicates high expression level. (CGAP website, n.d.)"
Does this mean that even if B>A it isn't automatically an over-expressed gene? So does it need to be red in the cancer section in th###we table for pancreatic cancer AND B>A?
What is the Figure 3? I guess user is talking about ‘Digital Northern Results’ page. Blue indicates less expression level, red indicates high expression level. In this page user can see the list of libs with colors, which can be in or Pool A, or Pool B. We don’t talking about A>B or A<B in this page. The color indicates the expression level of this tag, or of the gene mapping this tag, in this specific lab.

5.  How about if the normal tissue is also red? Does that mean that the gene is not over-expressed?
The color indicates the expression level of this tag, or of the gene mapping this tag, in this specific lab. Does not related with normal or cancer. The normal and cancer are related with lib. For example at the page http://cgap.nci.nih.gov/SAGE/FreqsOfTag ... ACAACCATCT normal lib
LSAGE_Embryonic###l_stem_cell_HSF6_normal_p50_CL_SHES9 has high expression level for the tag GTGGCTCACAACCATCT.

6.  Does the normal tissue need to be blue? No, see the above

7.  What does it mean if its pink or light blue? It says something about tags. Could you please explain tags?
It says the freq of the tags. See the above.

8.  Also, what can be figured out from over-expressed genes?
For example, one cure for cancer.
No answer.

9. Can something be achieved in finding genes that are common to two types (chosen from Carcinoma, Sarcoma, Lymphoma and leukemia, Germ cell tumor, Blastic tumor or blastoma OR the different kinds of tissue) of cancer?
No answer.

10. Can the over-expressed genes be "turned off"?
(this after the queries are submitted and the results are shown)
No answer.

11. It seems like it is the list of over-expressed genes, but what is it exactly?
No answer.