How excactly is the Q30 is calculated?

Question

[rodrigo]

Seven days ago we started a new sequencing run on our G400 sequencer from MGI. However, this time there seems to be an issue with the run time. The time the sequencer spends taking the image for each cycle is unusually long, which is why the run is taking much longer than expected. The G400 provides a real-time raw Q30 value, and to our surprise, the Q30 is still high (around 90).

My question is: could anyone explain how the Q30 is calculated? I know that fluorescence intensity and signal separation are taken into account, but is it really possible for the Q30 to remain this high despite the delay?

We will check the data very carefully once the run is completed, but any ideas or suggestions on how to approach this issue would be greatly appreciated.

Answer 1

[GenoMax]

Phred quality scores have been in use since the 1990s and early days of human genome project. They were co-opted in the fastq format as NGS technologies came online.

the Q30 is still high (around 90).

I assume you mean that 90% of bases continued to be over Q30. If so that is not unexpected with improvements in sequencing reagents/technology.

but is it really possible for the Q30 to remain this high despite the delay?

Current gen sequencing machines have powerful computers attached to them that can multitask (take care of basecalling while the sequencing and image data collection goes on etc). So the calculation of the Q30 should have no impact from the actual sequencing, since the image data can be buffered to memory, disk etc.

MGI is also using a similar "probability of incorrect basecall" to define the Q scores --> https://en.mgi-tech.com/Home/Applications/index/id/13.html#:~:text=By%20adopting%20the%20DNBSEQTM,is%20the%20Q%20quality%20score?