HyPhy message board - Print Page

HyPhy message board
http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl Methodology Questions >> How to >> Mean substitution rates http://www.hyphy.org/cgi-bin/hyphy_forums/YaBB.pl?num=1271196936 Message started by sspielman on Apr 13^th, 2010 at 3:15pm

Title: Mean substitution rates
Post by sspielman on Apr 13^th, 2010 at 3:15pm

Hello,
I am trying to find the mean substitution rate for particular regions in a gene using HKY85 model, but thus far have only been able to find per-site values for substitution rates using siterates.bf, and inputing my data, defining a partition, providing HyPhy with a tree, etc. returns only a tv/ts value.
Any suggestions would be much appreciated!
-Stephanie

Title: Re: Mean substitution rates
Post by Sergei on Apr 13^th, 2010 at 5:10pm

Hi Stephanie,

What do mean by an average rate? HyPhy will estimate branch lengths for you (measured as expected subs/unit time), but and if you want the actual rate, you need to estimate the evolutionary time by other means (fossil record, molecular clock etc). Could you elaborate on the problem you are trying to solve?

Sergei

PS You may find section 2.2. of Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to

Title: Re: Mean substitution rates
Post by sspielman on Apr 14^th, 2010 at 12:04pm

Essentially, I have a gene with several regions which are under different functional constraints, so sites within a region should have similar substitution rates. For instance, if sites 1-50 are in region 1 and 51-75 are in region 2, I am looking to find the mean of the per-site substitution rates within region 1 and again for region 2.
Is there a way to find the confidence interval or standard deviation for the individual site rates returned by SiteRates.bf that would allow me to calculate the mean with error values?

Title: Re: Mean substitution rates
Post by Sergei on Apr 15^th, 2010 at 9:26pm

Hi Stephanie,

Do you have time information, i.e. are you talking about actual rates (e.g. subs/site/MY), or simply genetic distances (subs/site/unit time)?

In case it is the latter, then all you need is to perform a relative-ratio type analysis for your partitions. Averaging site-by-site estimates is not a good idea (mostly because they will have vastly differing variances), I would recommend simply obtaining a mean (per partition) estimate of substitution rates. Take a look at Multimedia File Viewing and Clickable Links are available for Registered Members only!! You need to

Login

This will allow you to estimate the ratio between substitution rates in your multiple regions.

To convert this to actual rates, you need more information (e.g. divergence time estimates)

HTH,
Sergei