Pooling reads from different libraries

[jg3197]

Hi all,

I have a library I have been given to work with but am having trouble working out if my approach is legitimate. I have 3 libraries each containing 2 tissues from males, and another three libraries for females, one containing both tissues, and the other two with the tissues samples separately, e.g.

Sample 1 - Male - Tissue 1+2
Sample 2 - Male - Tissue 1+2
Sample 3 - Male - Tissue 1+2
Sample 4 - Female - Tissue 1+2
Sample 5 - Female - Tissue 1
Sample 6 - Female - Tissue 2

Is it legitimate to combine sample 5 and 6 to have two replicates for females to compare against the males? I wondered if this would cause problems for fpkms (which are quite sensitive to library sizes) and/or counts as the combined sample would then contain very wildly different numbers of reads?

Many thanks,

Jen

[jg3197]

I am still struggling with this problem and wondered if anyone had a similar experience? To clarify my problem further, the libraries detailed above were already sequenced before I got the data to work on, so I had no part in the experimental design, which I know is not ideal. So I actually have fastq reads. So, is it ok to combine two fastq files (for samples 5+6) to get one set of reads for tissues 1+2 which will be approximately double the sequence depth of all other tissue samples? The samples were treated in exactly the same way, except that they were separated before libraries were generated.

I think it is ok, but i'm not sure if there's something I have overlooked.

I hope somebody can help me! Thanks.

[fanli]

Are the 3 male samples biological replicates from 3 individual males? And similarly for the females? If so, then technically you'd be pooling variation between two females. Read depth shouldn't be an issue if you normalize anyways.

I'd probably do some quick QC with all 6 samples separately first (e.g. just see where everything falls on a PCA).

[jg3197]

Hi Fanli, thanks for your reply!

Yes, the males are all separate individuals. For the females, sample 4 is from one individual, and sample 5 and 6 are different tissues of the same individual. So in total, there are 3 males and 2 females.

I did a PCA, and samples 5 and 6 were quite different from each other, which is what i'd expect. I haven't tried a PCA yet combining 5+6 into one sample yet though.