increase in quality over the first 14 in illumina data

[Cris]

Hi,

I am new here and I coudn't find the answer in the forun, so here it goes: As far as I know the quality of the sequencing is decreasing by the end of the reads, but in my data I have a increasing quality until the 14th base, than it starts to drop.



Why is the quality increasing? There is a explanation for this to happen? If you could point me to a reference that explains this would be great.
Thank you very much!

[kmcarr]

http://seqanswers.com/forums/showthread.php?t=22939

[Vinz]

Yes, that is normal.
Why?
To determine Q values, they analyze some characteristics of the basecalls. With those characteristics they go into a table to make a Q value out of it. This table is derived from training data sets where they know the sequence. Now, for some reason they found out, that it works best, if they use different look up tables depending on the cycle number. That results in the strange picture of increasing Q values at the beginning of the read. Basically, the used look up tables max out at different Qs.
So far for a biologists explanation of whats going on...

[Cris]

Thank you Vinz!
Do you have any source or reference that I can go to use on a presentation? This is a question that we are debating with other problems relating to a sequencing.

Thank you again!

[Vinz]

Maybe something is mentioned here:
RTA Training video

And there is a document called "RTATheoryofOps" which describes lots of theoretical details about phasing, matrix, base calling... I can not find it online anymore. But your Illumina support should be able to give it to you.