3
$\begingroup$

Checking out on Google Scholar, I can see that for Illumina (just to consider one example) the sequencing error rate is of the order of 0.001-0.01 per nucleotide.

Talking about sequencing error, let's consider mismatches (substitution of one nucleotide by another) only. Knowing the "true" nucleotide at a given position, is it as likely to be read as any other specific nucleotide during a mismatch or are there bias? For example, if the true nucleotide is A, is it more likely to be found being as a G (as they both are purines) than a T or a C? Are some nucleotides more likely to be misread than others?

I am hoping the answer won't depend too much on the sequencing techniques.

$\endgroup$
  • 1
    $\begingroup$ Actually sailorgyaru's answer is very close. The base-calling in Illumina as is different from that of Sanger sequencing. In the latter a noisy position can be assigned to different nucleotides with different probabilities. In Illumina the bad quality can be because of multiple reasons- improper washing , bad signal capture etc. Poor washing may cause overall bad quality everywhere. Perhaps you may look at the mismatch frequencies of nucleotides in reads with overall good quality (mean phred >20). I suspect the mismatch is likely to be random. $\endgroup$ – WYSIWYG Aug 14 '15 at 4:38
1
$\begingroup$

Unfortunately, it does depend on sequencing techniques.

For example, in Illumina sequencing, each sequence fragment is amplified (in order to get a stronger signal) and forms a cluster on the microarray. Each cluster is sequenced by cycles of:

  1. Adding fluorescent terminator nucleotides. These nucleotides are modified to contain an inhibiting/terminating group and prevents more nucleotides from being added. Theoretically, only one nucleotide is incorporated into every DNA fragment in this step.
  2. Washing off excess nucleotides.
  3. Capturing the incorporated nucleotide using imaging techniques and determining which base was incorporated (based on the colour of fluorescence).
  4. Cleaving the terminator from the added nucleotides, so that the reaction can continue.

Procedure of Illumina sequencing

Image from Metzker, 2010.

This way, each fragment is synthesized, one nucleotide at a time, and each nucleotide that is incorporated gets detected. However, the first step is not flawless: sometimes more than one nucleotide gets incorporated into a certain DNA fragment, or no nucleotides get incorporated. Eventually, the DNA fragments in a cluster (all containing the same sequence) will fall out of sync ("phasing") and the fluorescent signal will become less clear, with a mixture of different colours. This is the main cause of sequencing error for Illumina machines, and also the reason why Illumina reads are relatively short (~300bp).

So to answer your question, in this example, nucleotides may be erroneously read as nearby nucleotides in that sequence. Errors will vary using other sequencing methods and how those methods work.

The article I linked earlier explains various sequencing methods in more detail. (Unfortunately, it's behind a paywall so some may not be able to view it.)

|improve this answer|||||
$\endgroup$
  • $\begingroup$ Welcome to Biology.SE and thank you for your answer. So just to make sure, you that indeed the probability for a A to be read as a given nucleotide is dependent of the nucleotide and dependent on the technic used? You're not only saying that the error rate is dependent on the technic, right? $\endgroup$ – Remi.b Aug 14 '15 at 2:25
  • 1
    $\begingroup$ I can't say for sure but I don't think it's dependent on the nucleotide (at least not the chemical properties of the nucleotide). Most sequencing methods use different coloured fluorescent markers to detect which nucleotide is incorporated, and which base is called depends on the intensity of the colour signal. Perhaps certain bases are more likely to be erroneously incorporated into DNA in general, but I've never seen that mentioned as a significant factor. $\endgroup$ – sailor mercury Aug 14 '15 at 2:30
  • $\begingroup$ In the future, we recommend linking to cited works via PubMed or DOI. Publisher's websites can (and do) change. Additionally, please double-check your links before posting. The link to the Metzker paper was broken. $\endgroup$ – MattDMo Aug 14 '15 at 4:03
  • $\begingroup$ @MattDMo Sorry about that; thanks for fixing it! $\endgroup$ – sailor mercury Aug 14 '15 at 16:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.