Problems with TopHat and Paired End RNASeq

[quique_vzquez]

Hi, I'm very new at RNASeq Analysis and I think that I'm doing something wrong.
I've tried to align that data: http://trace.ncbi.nlm.nih.gov/Traces...mple=SRS167019.

First I execute sratoolkit to extract the fastq files with this command:

fastq-dump --split-3 ./1_SRR097897.sra

Then I have two files: 1_SRR097897_1.fastq and 1_SRR097897_2.fastq.

The next step that I've done is to execute TopHat to generate the bam file with the reads aligned. The command was:

../tophat2 -N 0 Pombe 1_SRR097897_1.fastq 1_SRR097897_2.fastq

Then I use the IGB to visualize the results and than shows the + and the - strands are the same



If I only use one of the _n files to generate the output I have two strands with polarity, but depending on the file to select the polarity is reversed.



I have also tried running TopHat with different parameters enabling --library-type or a GFF annotation file with identical results.
I could stay only with the results obtained through _2.fastq file but I wonder to know what I'm doing wrong, for other occasions.

Thanks for your time!!

[rflrob]

I think it makes sense that the different paired ends will be on opposite strands (since you basically read the second end forwards, but on the other strand). At the level of zoom you're at in the included figures, I'm not sure that you could count on seeing the different spacing of the read1s vs the read2s (in my experiments, something like 100-300 bp). It doesn't seem like the best solution, but you could perhaps go through the BAM file and change all of the "SEQ being reverse complemented" bits in the FLAG field to get both ends pointing the same way.

[quique_vzquez]

Thanks for your answer.
Finaly I change the reads in one of the FASTQ files to their reverse complementary form and I can align them with polarity but I haven't understood yet why it works :S