HyPhy message board - re: constant rate model

	Welcome, Guest. Please Login

Home

Help

HyPhy message board › Methodology Questions › How to › re: constant rate model

(Moderators: Sergei, Simon)

‹ Previous Topic | Next Topic ›

Pages: 1

Send Topic

re: constant rate model (Read 2911 times)

bioinfo_ucc

YaBB Newbies

Offline

Feed your monkey!

Posts: 5

re: constant rate model
Aug 16^th, 2011 at 8:28am

Dear Sergei,

I have just started to use Hyphy in the hope that it will help me in my project. My problem seems pretty simple or atleast to you, should be simple.

I have an alignment of sequences and I have certain reasons to believe that there is some region in the middle of the alignment which has a regulatory role, hence should show enhanced conservation. And decreased synonymous (or even non-synonymous) substitution rate.
I saw your MBE paper (2005- Site-to-Site Variation of Synonymous Substitution Rates) which incorporates quite complicated and parameter rich models.

However I was thinking about a relatively simple approach. I could simply use the "Constant rates model". This would be my null hypothesis i.e. all sites in an alignment evolve under constant synonymous rates. Then I should have a probability of each site (codon) fitting the null model. And the outliers, that do not fit the model should be regions which, I believe, play regulatory role.
Do you agree or do I have to start with a Dual model to get to this conclusion ?

Regards

Sankalp

IP Logged

Sergei

YaBB Administrator

Offline

Datamonkeys are forever...

Posts: 1658
UCSD
Gender: male

Re: constant rate model
Reply #1 - Aug 18^th, 2011 at 9:52am

Hi Sankalp,

I would actually suggest that you use Datamonkey.org and one of the methods it implements to estimate dN/dS at individual sites (e.g. FEL or MEME) and then, as an exploratory step, just plot the estimates over sites and see if your expectations are borne out. The Datamonkey tutorial at Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to

Login is a good place to start.

Alternatively, you could use something that was called a fixed sites model, dating back to this paper Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to

Login

Basically, you partition your alignment into regions a priori and then test if average dN/dS differ between them. You can implement that pretty easily in HyPhy through the GUI, e.g. see Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to

Associate Professor
Division of Infectious Diseases
Division of Biomedical Informatics
School of Medicine
University of California San Diego

WWW

IP Logged

bioinfo_ucc YaBB Newbies Offline Feed your monkey! Posts: 5	Re: constant rate model Reply #2 - Aug 22^nd, 2011 at 1:58am Thank you very much Sergei. Regards Sankalp
Back to top	IP Logged

bioinfo_ucc

YaBB Newbies

Offline

Feed your monkey!

Posts: 5

Re: constant rate model
Reply #3 - Aug 22^nd, 2011 at 4:05pm

Hi Sergei,

I think I was wrong in my presumption. In my last post, I had mentioned how I want to figure out regions within an alignment of protein coding genes with regulatory function by trying to identify sites which have a reduced dS and dN.

However, reduced dS and dN simply imply that such sites are well conserved and not tolerable to any sort of substitutions. This inturn implies that the protein sequence is important, not necessarily the nucleotide sequence.

If however, we find sites with low dS, irrespective of what dN is for those sites, it would imply that the selection is at the level of nucleotide and not peptide sequence.
I have seen a few papers where people have tried to identify Exonic splicing enhancers by identifying regions with low dS, considering that ESE is another regulatory element in protein coding regions.
Do you agree ?

Regards
Sankalp

IP Logged

Sergei YaBB Administrator Offline Datamonkeys are forever... Posts: 1658 UCSD Gender:	Re: constant rate model Reply #4 - Aug 25^th, 2011 at 3:19pm Hi Sankalp, Sure, you can look for reduced dS as a proxy for nucleotide level conservation (with invariable sites being the complete extreme). You should still be able to explore this using FEL or REL on DataMonkey and plotting dS (or E[S]) over sites. Sergei
Back to top	Associate Professor Division of Infectious Diseases Division of Biomedical Informatics School of Medicine University of California San Diego WWW IP Logged

Pages: 1

Send Topic

‹ Previous Topic | Next Topic ›

« Home

‹ Board

Top of this page