Reply
Highlighted

Fastq read filtering prior to supernova run

Posted By: jpenalba, on Jun 3, 2017 at 3:12 AM

 Hi,

 

I was wondering what kind of read filtering and trimming would be necessary prior to running supernova run. Alluding to another question in the forum, is this integrated within the pipeline or something we should run separately?

- Remove low quality reads

- Remove duplicate reads

- Trim adapter sequences

- Remove low complexity reads

 

All thoughts and experience are appreciated. 

 

Best,

Josh

3 Replies

Re: Fastq read filtering prior to supernova run

Posted By: melop, on Jul 3, 2017 at 6:45 AM

From my understanding the supernova pipeline includes a quality control step. The idea is that you can feed the raw data into the pipeline and it produces a great genome. 

 

My run has some issues on read 2 quality, I'm actually wondering about the same thing. But I guess we just need to try and compare.

Re: Fastq read filtering prior to supernova run

Posted By: graham, on Sep 15, 2017 at 6:19 AM

Hi Josh,

I've not played around with read trimming as Supernova does its own trimming step, but I'd point out it would be a bad idea to do 5' trimming on R1 reads as these contains the barcodes. Trimming barcodes could lead to a whole host of problems!

Cheers,

Graham

Dr. Graham Etherington
The Earlham Institute, Norwich, UK
Twitter: @bioinformatiks

Re: Fastq read filtering prior to supernova run

Posted By: melop, on Nov 6, 2017 at 4:29 AM

Now I have tried trimming adapters (but not by quality). First, the pipeline has some problem with trimmed reads. Make sure all reads < 100bp after trimming are thrown out, otherwise the pipeline may crash. 

 

The result looks worse than using the raw data, consistent with what the 10x folks told me.