RNAseq merging runs

[quinne5]

Hi,

I was just wondering if anyone had any comments on whether there are advantages/disadvantages to merging fastq files from different sequencing runs of the same sample over merging the BAM files with regards alignment(using Tophat) ?

Also -for some of our samples we had sequencing runs with low read numbers 4-8 million reads (single end) & were advised not to merge these in with later runs for the same sample that had higher numbers so I was also wondering if anyone had any comments on that either as I'm not really sure why this is so.

Thanks in advance for any advice!!

[mimamo]

I am also wondering about this.
I have RNAseq samples run over four lanes with HiSeq and thought I'd use TopHat for alignment. Should I merge the samples before or after running TopHat? It seems like TopHat is slow enough running just one seq run at a time, but I am afraid that I'll loose some information if I don't align what has been run on different lanes into one file.

[Richard Barker]

You can analyze your files separately and then merge your repeats during the cuffdiff stage, the iplant website has a really user friendly tutorial and the iplant collaborative cloud infrastructure is really useful as you can perform all your analysis without freezing up your computer

Iplant general info

http://www.iplantcollaborative.org/

Iplant RNA-seq_tutorial (with cuffdiff recommendation of repeat merger point)

https://pods.iplantcollaborative.org/wiki/display/eot/RNA-Seq_tutorial