What's the effect of large difference of library size by RNA-seq ?

[yeyeming]

We RNA-seq two samples from the same tissue but different groups using SOLID,and the number of reads are 28 million and 95 million, is this normal,and what is the effect of such large difference?

[pmiguel]

Did you do the sequencing yourself? If so could you tell us what the enrichment % was for each library? Were these barcoded and put into the same region, or in different regions?

It is possible, for many reasons, to see large differences in the numbers of sequences from similar samples.

--
Phillip

[yeyeming]

Did you do the sequencing yourself? If so could you tell us what the enrichment % was for each library? Were these barcoded and put into the same region, or in different regions?

It is possible, for many reasons, to see large differences in the numbers of sequences from similar samples.

--
Phillip

Thank you for your reply,a pity,we did not do the sequencing by ourselves,what i got are just the csfasta and qual files of each library.so i have no idea about the enrichment% and the mean of "Were these barcoded and put into the same region, or in different regions".
Can i get some indexes from the csfasta and qual files to decide whether i could go on my analysis? thanks.

Ye

[pmiguel]

Hi Ye,
My advice would be to ask the facility that did the sequencing for you why the numbers are so different for the samples.
Have you mapped the sequences against your reference genome? If so what were the mapping percentages and what organism was it?
You might also want to plot a histogram of quality scores for both data sets. That would give you an indication of whether there are vast differences in their sequence quality.

--
Phillip

[dzavallo]

Hi Ye, I have the same problem between my 3 treatments: 10, 6.4 and 5.7 millon on each. And when I mapped against the reference genome the percentages were very different too.
How did you normalized your counts?

[ishmael]

I agree with Philipp, many factors may effect final sequencing result.
Were the datasets sequenced in same batch?
What the results of FastQC?
You may check the top 20 expressed sequences of each dataset, and they may give you some clues.

[yeyeming]

What's tool you use to map? I use Tophat to analysis,the accepted hits of my two libraries acount for 42% and 57%,and then use cuffdiff (in cufflink )with FPKM to do different expression analysis.

[CC_seqanswers]

It's normal.

The two groups might just have different number of libraries running on a flowcell or library normalization was off before emPCR. The number of reads your sample has depends on the proportion of your library in the whole sequencing pool. Ask whoever handles the agreement between your group and the sequencing service provider to see how many raw reads you are supposed to get based on what you paid for. You might be just lucky to get 95M reads as you are supposed to get only 28M. Or maybe it's' the other way around.

We RNA-seq two samples from the same tissue but different groups using SOLID,and the number of reads are 28 million and 95 million, is this normal,and what is the effect of such large difference?

[Marianna85]

Hi Ye, I have the same problem between my 3 treatments: 10, 6.4 and 5.7 millon on each. And when I mapped against the reference genome the percentages were very different too.
How did you normalized your counts?

Hi dzvallo,
I'm dealing with a similar problem...at the end what did you do? how did you perform the normalization?