trimming in tophat

[frymor]

Hi all,

I am trying to analyse my PE Illumina data using tophat.

At first I run fastqc. Checking the raw data, I discovered at the beginnings (and presumably at the ends) of my reads I have some containments from the adapters of the sequencing.
I run bowtie first on both the full length and trimmed sequences and got better results with the trimmed sequences.

Do I need to trim the data before running tophat?

Does someone know how to do it? do I need to convert my trimmed sam files (bowtie output) back into fastq files?

Thanks for any help
Assa

[dnusol]

Hi, I found this useful page about this issue.

http://bioinfo-core.org/index.php/9th_Discussion-28_October_2010

HTH

Dave

[frymor]

Thanks for the tip.
It is a good page with summaries about the different images of the fastqc software, a thing a lot of people were looking for in a different thread.

BTW, can anyone tell me of a good way to remove the duplicates reads from the equation.
running the fastqc program I get a lot of duplicated reads (see attachment).

As I am looking for differentially regulated genes I am not sure whether I should exclude the duplicated reads or not, but I would like to try ans see what I get when doing so.

Q: can anyone tell me how to filter duplicated genes from the sam files or before the bowtie run from the fastq files?

Q: Is it the right way when going for differential expression also to exclude the duplications? or do I need to keep them?

Thanks

Assa

[jp.]

did you get the answer ?
would like to share it here
thank you

Thanks for the tip.
It is a good page with summaries about the different images of the fastqc software, a thing a lot of people were looking for in a different thread.

BTW, can anyone tell me of a good way to remove the duplicates reads from the equation.
running the fastqc program I get a lot of duplicated reads (see attachment).

As I am looking for differentially regulated genes I am not sure whether I should exclude the duplicated reads or not, but I would like to try ans see what I get when doing so.

Q: can anyone tell me how to filter duplicated genes from the sam files or before the bowtie run from the fastq files?

Q: Is it the right way when going for differential expression also to exclude the duplications? or do I need to keep them?

Thanks

Assa

[frymor]

No I didn't get any response for the questions I posted.

I am not sure though how important is the duplication rate in this step. I'm using tophat2 with the option to exclude all duplicated reads, so I am not worried about the duplication in the original fastq file.

I hope I am thinking in the right direction.

[anamika]

Sangenix

SangeniX: A comprehensive, automated, scalable and user friendly NGS data analysis suite

Sangenix Has module for duplication removal.

Give it a try : http://www.sangenix.com/

[frymor]

let me know again, when it is a freeware

[vineet jha]

Sangenix

Beta Version is available. you can contact to us via contact page in http://www.sangenix.com/contactus.aspx

[dpryan]

Removing the duplicates could be done with the samtools rmdup command (you could alternatively use markDuplicates from picard). This is generally not needed for RNAseq, since a certain amount of duplication would be both expected and desired for highly expressed genes (i.e., many/most of these probably aren't PCR duplicates).