Error rate of Illumina MiSeq

Question

[bioplanet]

Hi all,

I wanted to know if there is a study/paper where I can see the Illumina error rate for Miseq data.

I ran across this older post:

What is the illumina sequencing error rate?

and this paper:

http://onlinelibrary.wiley.com/doi/10.1111/j.1755-0998.2011.03024.x/full

from 2011, where it says that the error rate is 0.1% but only according to the company's report. Is anyone aware of a study that tell us more? My question is basically, if I have a sequence of 500 nts, can I expect a 0,1% error rate per position? Per the whole read? Something else?

[GenoMax]

bioplanet : It looks like you have not validated many of the answers for your past posts. Up-voting useful comments/accepting answers (green check mark) provides feedback from original poster for threads. It is also the appropriate way to show your appreciation for the help you receive on the forum.

[chen]

According to my experience, it's a bit higher than 0.1% for Illumina sequencers, probably 0.15% ~ 0.3%. And it varies by the sequencing conditions (i.e. cluster density).

[Kevin Blighe]

As per chen's comment, the error rate is not something that's easy to quantify. There are so many factors (many that we don't even fully understand) that will contribute to error. There are also different levels of error even in the sequencer:

• incorrect base addition on the growing strand by polymerase (for sequencing by synthesis)
• incorrect amplification by PCR
• incorrect base-calling on the fluorescent intensities

I have a colleague who worked a large part of her/his career on the development of the technology that would eventually be used by Illumina's sequencers. Illumina actually purchased the rights for this technology and did not develop their own [technology]. I don't know what sort of trade secrets s/he can divulge, and I would be reluctant to further divulge here on an online forum.

Edit: I believe that the quoted figure of 0.1% relates to the incorrect base addition by polymerase during the growth of the strand during sequencing by synthesis. These types of errors also occur in our own genomes during mitosis/meiosis, but we have proteins and signalling cascades to easily recognise these and repair them via base excision repair. In the sequencer, there is no such repair mechanism.

[GenoMax]

I don't think that error rate is per read. It should be for the total number of bases you have from run.

[bioplanet]

Thank you all for the answers.

So, basically, if I get it correctly, Illumina does not tell us that "for each given position there is a chance of 0.1% that I will read the wrong nucleotide", right? And, if this assumption is correct, then the 0.1% reported error rate cannot be used as a "base-line" number.

[GenoMax]

Please use ADD COMMENT/ADD REPLY when responding to existing posts to keep threads logically organized.

Submit Answer should only be used for answers for the original question.

[bioplanet]

Apologies for the mistake!

[Kevin Blighe]

As per genomax, as I also understand, it will generally mean that 0.1% of all bases from your run will be erroneous, and that's just the error from the sequencing step within the sequencer. That is, for every base that's sequenced, 0.1% will be erroneous.

This is not even considering the later bioinformatic steps, where mis-alignments can occur, etc.

[bioplanet]

Ok, many thanks to all of you for your time and responses!