How To Use Ests To Examine Gene Expression

Question

[Stephen]

A colleague of mine claims to have used ESTs to show that gene G is expressed in tissue T. How does one do this? Suppose I wanted to look at all the EST data available for a gene, say EGFR? What would I do?

[Larry_Parnell]

A well annotated EST entry will give info on the source library and this should then contain info on which tissue (or pooled tissues) were used to build the cDNA library that was then sequenced.

Answer 1

[Sean Davis]

Not an answer to do with ESTs, but this might answer your question using microarray from BioGPS.

http://biogps.org/#goto=genereport&id=1956

Of course, you can search for any gene of interest, not just EGFR.

Answer 2

[Neilfws]

This is not a very computational solution, but I would:

1. Search UniGene at NCBI for EGFR
2. Click the first link to EGFR for Homo sapiens
3. Scroll down to section GENE EXPRESSION and explore the links, especially the EST profile

Answer 3

[SES]

I don't work on human genetics, but I can tell you generally how people approach this question.

To say that a gene is expressed in tissue T you would need to isolate RNA from tissue T and construct a tissue-specific cDNA library. Usually the developmental stage is also important for where a gene is expressed.

To confirm your hypothesis about tissue specific expression, you could do in situ hybridization to demonstrate the timing/location of gene expression.

Answer 4

[Mary]

You can do a quick check on some using the UCSC Genome Browser. You can set the EST filter to contain a tissue name and be a certain color. However, that will only flag the ones that are annotated to the correct tissue. The biggest tissue type in GenBank is N/A.

Edit for afterthought: if you aren't restricted to ESTs, another tool I always thought was cool was the "digital northern" with SAGE: http://cgap.nci.nih.gov/SAGE/AnatomicViewer It will compare normal vs cancer tissue, but you can just mine the normals.

Answer 5

[Larry_Parnell]

In this paper from 1993, a human brain cDNA library was sequenced giving rise to some 3400 new expressed genes. A year earlier, the same group performed a proof of concept with ESTs from a mixed-stage library from C. elegans. Thus, knowing the source of the mRNA - brain - and carefully copying that mRNA into cDNA and sequencing in an unbiased manner, one can confidently state that the sequences detected are expressed in that tissue.