Non-Transcribed Strand And Transcribed Strand Mutations In Sequencing

Question

[Liyf]

I really do not know how a mutation only occur on non-transcribed strand or only on transcribed strand. I think non-transcribed strand and transcribed strand are complementary, so the number of mutations should be the same. Where are my errors?

[Leonor Palmeira]

You should give us the code you are using to compute your mutation rate, or at least the steps of your method, so we are able to help you.

[Liyf]

Sorry， but I got totally confused about this problem. Many paper reported this. I take one as example: Only T>C and C>A changes but not C>T changes were effectively repaired on the tran-scribed strand , and these repairs occurred more frequently in highly expressing genes

[Leonor Palmeira]

I have the feeling you might be talking about the CpG effect? If so, this paper might help you: http://www.ncbi.nlm.nih.gov/pubmed/11070050

[Liyf]

Thank you for your help! In fact I do not talk about CpG effect, what I am confused is mutations on transcribed strand and non-transcribed strand. Some reports say that because some DNA repair are specific for transcribed strand, so non-transcribed strand would have much more mutations. But I do not know how to count it just using exome sequencing data.

Answer 1

[Cliff Beall]

Mutations occur on one strand initially. To have a permanent change repair or replication needs to change the other strand. So the final mutations will have both strands changed. Repair operates differently depending on the nucleotide changed and on transcribed and nontranscribed strands (presumably because of RNA polymerase etc. coming through on the transcribed strand).

In the example you give, they are inferring the biased repair patterns from observed mutations. They are just not mentioning the complementary change on the nontranscribed strand for brevity and clarity.

[Liyf]

Thank you! You got my question! But I still do not know the example I give, is it just count the mutated reads number?

[Cliff Beall]

That is based on mutations detected and they are saying they see more T>C mutations on the untranscribed strand than the transcribed, for instance.

They are finding the mutations by aligning the sequence reads to the genome with BWA and using SAMtools pileup. So they aren't counting the reads really.

[Liyf]

Yes, I think you know what I am confused! Could you please explain more details? Samtools mpileup command seems not tell whether mutations are in transcribed strand or not.

[Cliff Beall]

It would have to be integrated with the genome annotation to know where the transcribed regions are and which strand is transcribed.

[Liyf]

Thanks. However, what I am confused is that how a mutation is just on one strand? Because both direction have mutations.

[Cliff Beall]

Yes, both strands have mutations. But one could count T>C mutations on one strand, and A>Gs as T>Cs on the other strand, which is how they are doing it.

[Liyf]

I got confused, would you please explain more detail? Or you could take an example?

Answer 2

[siddarth.dt]

http://www.ncbi.nlm.nih.gov/pubmed/9196330

This should answer your question