Ka/Ks Ratio To Detect Selection (Nonsynonmous To Synonymous) In Dnasp

hello everyone,

I have sequencing data for my gene from >2000 individuals. I want to look at the Ka/Ks ratio of my coding regions to assess evidence of selection.

I am aware that MEGA and DNAsp are good tools to find this out but I am confused as to what to provide the software.

I would prefer to use DNAsp as I know this better - can anyone answer the following?

1) DNAsp allows you too assign which bits of the input sequence are coding/non coding. I can give DNAsp the fasta sequences for each chromosome, but as my gene is huge it would be alot easier to just provide the coding sequence.

...so can I just put the coding sequence together as one long sequence in FASTA format, assign it ALL as coding and start my analysis?

It might seem a ridiculous question but I am just unsure of whether this will affect the analysis?!

Can anyone tell me? Or has anyone used this software for this type of analysis?

Thank you!!

• 7,561 views • link

updated 14.9 years ago by Casey Bergman

• written 14.9 years ago by User 7433 •

1 answer

To begin, if you have intraspecific polymorphism data, you should be calculating pi or theta (not Ka or Ks) for synonymous and non-synonymous sites.

To answer your main question, unless your alignment is so large that it cannot be read into DNAsp natively, there is no reason to extract coding sequences since as you are aware DNAsp allows you to input a set of functional sites that annotate the region for coding and noncoding regions. In theory, these approaches should be identical, barring the summarization of silent sites which in a genomic context will be for synonymous + noncoding sites, while in a CDS-only context will be just for synonymous sites. Additionally, unless you have for example sequenced only exomes, you will have to do as much or more manipulation to extract the CDS component of your alignment as you would need to do to input you CDS annotation to DNAsp.

• 1 views • link

written 14.9 years ago by Casey Bergman

Hi, thanks for your answer.

I have calculated nucleotide diversity (pi) and gene diversity (h) but am still interested in the Ka/Ks ratio.

So it should not make a difference if I put the intronic region between exons into the sequence or not. As technically these will be excluded anyway once I assign the coding regions in DNAsp..

okay thank you! x

• 0 views • link

written 14.9 years ago by User 7433 •