Digital gene expression count from RNA-seq??

[ritzriya]

Dear All,


1) I have been working to code for a tool to get the digital gene expression count for genes across different samples. I wanted to know that can the Digital gene expression level be taken as the number of reads mapping to that gene?

If yes, then is that an accurate way of estimation of getting the expression level without going for a separate experiment to find the levels?

2) Also, I have heard that using cufflinks coverage, we can derive at a count for a gene - is that true? I found it confusing as the coverage values provided by the software, is observed always in decimals as 0.007 or may be 1.05. How can a gene be expressed in such terms? Am i missing some logic here?

Will be glad to receive some answers for the above queries.

Thanks in advance!

[ritzriya]

Has this been repeated in the forum?

***************************
I wonder if this question has been answered in some other thread! I am sorry if I did so, but can anyone give a quick link to it if yes?

If no, how come noone has the answer in SEQanswers!!? Strange..

Please help.. I need it asap..

Dear All,


1) I have been working to code for a tool to get the digital gene expression count for genes across different samples. I wanted to know that can the Digital gene expression level be taken as the number of reads mapping to that gene?

If yes, then is that an accurate way of estimation of getting the expression level without going for a separate experiment to find the levels?

2) Also, I have heard that using cufflinks coverage, we can derive at a count for a gene - is that true? I found it confusing as the coverage values provided by the software, is observed always in decimals as 0.007 or may be 1.05. How can a gene be expressed in such terms? Am i missing some logic here?

Will be glad to receive some answers for the above queries.

Thanks in advance!

[Simon Anders]

***************************
I wonder if this question has been answered in some other thread! I am sorry if I did so, but can anyone give a quick link to it if yes?
If no, how come noone has the answer in SEQanswers!!? Strange..
Please help.. I need it asap..

Maybe you are a little bit impatient.

If you want to count the number of reads falling onto a gene, maybe have a look at my HTSeq framework, especially the htseq-count schript.

Whether this is a good measure for the gene expression depends on what you mean by the latter.

- If you want to compare one gene across several samples, normalization is crucial, but simple. See the section on normalization in our paper on DESeq or Robinson and Oshlack's paper on the subject.

- If you want to compare the expression of different genes within the same sample, you need to divide by the transcript length, and for this, you need to know which transcript is expressed. This is discussed in the Cufflinks paper.

Simon

[ritzriya]

Thank you Simon for the reply! Yes I guess I was being a little impatient...oops..

Well, I shall have a look at the links you have provided.. thank you again..