units of divergence time estimates

[Ismail]

Hi all,

I am a new user and would like to clarify whether my
estimateddivergence times have the correct units (i.e on a per million
year bases).

I estimated divergence times for a closely related set of species
however since I had no prior fossil information I calibrated the tree
by putting a prior distribution on the mean substitution rate as
suggested here : http://groups.google.com/group/beast-users/browse_thread/thread/9719927b4d9aabe7.

I had a couple of independant estimates of substitution rates from
other studies with a mean of 2.04 ± 0.06 % per million years. As
suggested by "Alexei" I fitted a lognormal distribution to the rate
estimates obtained from other studies which gave me a mean of 0.67 ±
0.29. I fitted the lognormal distribution to the raw values and did
not scale them (i.e. I used 2.05 not 0.0205 while fitting the
lognormal distribution). I then later used these estimates to describe
the log normal prior distribution for mean substitution rates.

My results were as follows: MeanRate = 2.06 TMRCA = 0.03942. I am
assuming that the unit here is "million years" i.e TMRCA = 39420 years
(roughly 40 thousand years).

I would be very happy if someone could clarify if my units are correct

cheers
Ismail

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To post to this group, send email to beast...@googlegroups.com.
To unsubscribe from this group, send email to beast-users...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/beast-users?hl=en.

[paolo gratton]

Hi Ismail,

It looks that your time units are not actually million years.

You set the prior on the rate in units of % per million years, instead
of substitutions per site (per lineage) per million years.
Therefore, your actual unit is 100 Myr, and TMRCA = 0.03942 should
ltranslate 3.942 milllion years.

Hope it may help.

P.s.: if you are analyzing mtDNA data, 2% is a typically reported
DIVERGENCE rate, did you check you are instead using reported
SUBSTITUTION rates (per lineage, and thus = divergence rate/2)?

I am interested in substitution rates in invertebrate's mtDNA, so, if
you are studying this kind of data, I would be really pleased if you
could send me the references for the 2.04 rate you are using (write me
at paolo.gratton at gmail.com). Thanks a lot.

Best

Paolo



2010/4/19 Ismail <iksa...@gmail.com>:

[julien]

Hi Ismail,

I totally agree with Paolo

Something I wrote elsewhere but did not appear... was about setting
logmean and logstdv for lognormal prior... It may be useful for your
case study...

Given the 95% CI you want... you just need solve this set of simple
equations to get the corresponding mean and stdv (standard deviation)
in log space:

upperbound= exp(mean + 1.96 stdv)
lowerbound= exp(mean - 1.96 stdv)

This allows you to set accurately what you want.

So if you solve these equation assuming a lower bound of 0.0198 (1.98
%) and an upper bound of 0.021 (2.1%), then you should set a lognormal
prior on meanrate (or ucld) with:

log(mean): -3.892653091
log(stdv): 0.015010332

Best

Julien