Tag counts for RNA seq experiment

[sanush]

Hi all
We have to design an experiment to sequence the transcriptomes of 2 different cell types to know what genes are differentially expressed.

ABI recommends different Tag numbers (mappable reads) for different applications but they don't specify the exact (atleast approximate) mappable reads required to survey for different applications.

So to study the differentially expressed genes in different cell types what is the recommended mappable reads or tags we need to get a good coverage.

Thanks for the suggestions

[pmiguel]

Hi all
We have to design an experiment to sequence the transcriptomes of 2 different cell types to know what genes are differentially expressed.

ABI recommends different Tag numbers (mappable reads) for different applications but they don't specify the exact (atleast approximate) mappable reads required to survey for different applications.

So to study the differentially expressed genes in different cell types what is the recommended mappable reads or tags we need to get a good coverage.

Thanks for the suggestions

In case you missed my answer in the other thread you started:

http://seqanswers.com/forums/showpost.php?p=11053&postcount=7

Here is what I posted:

Brian Coullahan, Lifetech Application Specialist, gave a talk at the 2009 SOLiD Summit in September. He recommended:

SREK 10-20 million mapped reads
SWTAK 10-100 million
SOLiD SAGE 2-5 million

--
Phillip

[sanush]

Thanks Phillip
I have that slide too, but it gives us a range 10 - 100 M. Is there any specifics like the number of mappable reads required for different applications... what difference can we see in final data with lower reads and higher reads.

My question is should we have to aim for higher reads(90M) or around 30-40M reads, because as we go for high reads then we need to load 1 sample in multiple segments in quad slide which could increase the cost of the experiment.

Really appreciate your suggestion

Subu

[pmiguel]

Thanks Phillip
I have that slide too, but it gives us a range 10 - 100 M. Is there any specifics like the number of mappable reads required for different applications... what difference can we see in final data with lower reads and higher reads.

My question is should we have to aim for higher reads(90M) or around 30-40M reads, because as we go for high reads then we need to load 1 sample in multiple segments in quad slide which could increase the cost of the experiment.

Really appreciate your suggestion

Subu

It is hard to say.
For polyA+ RNA 30M should be plenty unless you are focused on transcripts with low relative abundance (below, say, 10 transcripts per million), I would think. For non-polyA+ RNA (eg, ribominus) for which you also need information on poly-adenylated messages--then maybe you would want to get closer to 100M because structural RNAs will occupy a large proportion of your sequence space.

You can always add more sequence later if you don't have enough.

--
Phillip