Newbie needs help on a simple calculation

[kwatts59]

Hi all.
I have been reading up on Phred quality calculations.

Phred quality = -10log(p) and p is the probability the base is NOT correct.

My mentor wants to know how p was calculated and wants a working example. I searched high and low and I cant find a working example.

If someone can show me an example of how to calculate p, it would be greatly appreciated.
Thanks in advance

[lmf_bill]

http://en.wikipedia.org/wiki/Phred_quality_score,
maybe, you can get the answer.

[d17]

For the original Phred scores you can see how p is calculated from the papers describing the method:

http://www.ncbi.nlm.nih.gov/pubmed/9521921

http://www.ncbi.nlm.nih.gov/pubmed/9521922

However this is for Sanger reads which I doubt you are talking about if you are posting on this forum. It is far less transparent how p is calculated for base qualities from Illumina, ABI SOLiD, etc. It will be dependent on the technology. Of course Phred scores are also being adapted in many other applications, e.g., NGS read mapping programs to measure the quality of an alignment. For such programs the details behind the calculation of p will be program-dependent.

[kwatts59]

It appears that people are not understanding my question.
I know how to calculate Q, the Phred quality.
The question is how to calculate "p".

http://s1091.photobucket.com/albums/i396/kwatts101/?action=view&current=Peaks.png

Let's say for sake of argument that the intensity of T is 1, the intensity of A is 0.5, and the intensity of G and C is 0.25.
What is the value of "p" that is used to calculate the Phred quality Q?

Q = -10log(p)

[d17]

My response above is answering your question. The answer is that how p is calculated will be program-dependent.

If you have Sanger reads that you are analyzing with Phred the second linked ref in my response above describes how error probabilities are calculated.