How many sequences for Ka/Ks Ratio

Question

[Charlotta]

Hello,

I wish to calculate Ka/Ks Ratio and test for selection with Mega. I have several sequences of a gene in several species. Do I enter all the sequences or should I only enter one sequence per species? In this case which sequence should I choose?

Thanks

Charlotta

Answer 1

[natasha.sernova]

See this posts:

Best Practices/Softwares To Calculate Ka/Ks Ratio - it's a nice guide what to do and how to deal with it.

Ka/Ks (Or Dn/Ds) Calculation From Sequenced Transcriptome Contigs - it's not your case, right?

Read this paper, I've selected a citation from it.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3055854/

“We choose human and eleven other high-coverage mammalian genome data”

It shows about the upper limit of information you can obtain.

Look at this site:

https://en.wikipedia.org/wiki/Ka/Ks_ratio

see Limitations there

The answers to your questions depend upon a set of your species and your goals.

You may want to calculate the ratio for one/several/many orthologous genes in all species, for example.

Probably you will need to build orthologous raws or groups to select data to study.

Use program PAML as a tool to calculate the ratio you need.

See this post, for example, A: dN/dS ratio (omega)

PAML and its manual are not pretty easy.

[Charlotta]

Thanks Natasha. I understand better. However I have still one question. I have one gene but several alleles for this gene in 10 species. Shall I use one allele per species (I don't have frequencies) or all alleles per species ? That's what is not clear for me.

Many thanks

Charlotta

Answer 2

[natasha.sernova]

This is an online calculator, it makes it faster:

http://services.cbu.uib.no/tools/kaks

You have one gene but several alleles for this gene in 10 species.

“Shall I use one allele per species (I don't have frequencies) or all alleles per species” ?

If you take one allele per species, how will you choose which one to select?

I would try to take all alleles per every species at first,

align them with the calculator above or somehow else – it will show you how different they are

and find allele consensus according to the post below.

Test of selection across whole genomes

Here Nicolas suggested to choose allele consensus.

In this post PAML Batch File? a pairwise topic is discussed. See @Brice advice in the end.

My impression is that there is no strict rule how to accomplish the task.

I would find a close target organism and study how people dealt with it before.

[natasha.sernova]

I found several paper for really different species:

Heterogeneous genetic diversity pattern in Plasmodium vivax genes encoding merozoite surface proteins (MSP) -7E, −7F and -7L

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4300842/

CODON-BASED DETECTION OF POSITIVE SELECTION CAN BE BIASED BY HETEROGENEOUS DISTRIBUTION OF POLAR AMINO ACIDS ALONG PROTEIN SEQUENCES

http://dambe.bio.uottawa.ca/publications/2006CSBXiaKumar.pdf

The KA/KS Ratio Test for Assessing the Protein-Coding Potential of Genomic Regions: An Empirical and Simulation Study

http://xungulab.com/publications/198.pdf

Investigating the Bird Flu Virus

http://www.mathworks.com/examples/bioinfo/mw/bioinfo_product-birdfludemo-investigating-the-bird-flu-virus

What is the organism you study? In general: human/mouse, viruses, Plasmodium, etc?