From: James Taylor <james@jamestaylor.org>;
To: Benjamin Osei-agyeman <benjy_osei@yahoo.co.uk>;
Cc: Ido Tamir <tamir@imp.ac.at>; galaxy-user <galaxy-user@lists.bx.psu.edu>;
Subject: Re: [galaxy-user] Quality report
Sent: Wed, Nov 6, 2013 6:05:26 PM

On Wed, Nov 6, 2013 at 12:54 PM, Benjamin Osei-agyeman <benjy_osei@yahoo.co.uk> wrote:

The report was obtained after running fastqc on my data. However after looking at the report statistics, the
Quality is not good in the Per Base Sequence Quality and there is a difference between the mean and median at each cycle. How can the quality be improved?

Is the quality poor across the entire read, or just at one end? You can improve overall quality in two ways. By filtering out lower quality reads (the Filter by Quality tool or Filter FASTQ tool) or by trimming a portion of the reads (the FASTQ trimmer tool).

It is important to understand that these tools work by throwing away data, you can't improve the overall quality of reads you already have. The FASTQC tool is telling you about the quality of your sequencing experiment. The only way to improve the overall quality is to do the experiment again.

And why is there a difference between the mean and median in the fastqc report at each cycle?

It means your quality score distribution is skewed. This is not necessarily a concern.

--
James Taylor, Associate Professor, Biology/CS, Emory University