Re: users of Phred

Robert Lyons (boblyons@umich.edu)
Thu, 15 Jul 1999 09:45:21 -0400

"Jacobs, Chris" wrote:
>
> I'd like to begin using Phred to analyze sequence data off our ABI 377's.
> I'm a serious beginner with this program and am a little daunted by it's
> command lines.
>
> Does anyone have suggestions?
> How do you score the results?
> I've heard people say 'a Phred score of 20' for example, what does that
> mean???
>
> Help!
> Chris Jacobs

Hi, Chris -

Your best bet at first is to just use the 'phredPhrap' script that Phil
supplies with the programs. It supplies all the proper command lines args
and sets up the entire dataset for import into consed. The phred Q score is
a programmatic assessment of the likelihood of error for the base under
scrutiny. Q = -10*log(error-probability), so a phred of 20 means there's a
10^2 (10 to the 2nd power) error probability (i.e a 1 in 100 chance). Data
destined for Genbank is supposed to have only a 1-in-10000 error rate,
corresponding to a Q of 40 (cumulative quality score calculated by phrap
during assembly).

Feel free to give me a call if you have further questions, Chris.

Bob Lyons
University of Michigan