substitution rate

[Heather E]

I am new to BEAST and have a very basic question.  I am trying to estimate divergence times using a gene with a rate of divergence approximately 4.5% per my, according to the literature.  Am I correct in that the rate of divergence can be used to set the molecular clock?  If so, when I put the rate into beauty, do I enter 4.5 or .045 as the clock rate?

[Huw A. Ogilvie]

Hi Heather,

It depends on exactly what the 4.5% refers to. It sounds like it might be the molecular divergence between two taxa after 1my, in which case the substitution rate will be half the divergence. In that case you should set the clock rate to 0.0225. Otherwise if refers to the substitution rate per my, then the clock rate should be set to 0.045.

- Huw

[shweta shukla]

As I have the substitution rate is 7.1X10^9, and generation time is in per year,  so how could I put these value to get the divergence time of my datasets ??

I have plenty confusion regarding the substitution rate and clock.rate.

[Huw A. Ogilvie]

Is 7.1x10^-9 the substitution rate per year or per generation? If per year, then you could set the clock rate to "7.1e-3" and your tree will be calibrated in millions of years. If per generation, you should divide the rate by generation time first to convert to per year.

The substitution rate is generally for estimating the relative rate among loci/partitions - this is the case when the "fix mean rate" option is set. It's possible to use this parameter to estimate the absolute rate, but in my own analyses I prefer to use it for relative rates, and then use the clock rate to set the mean absolute rate among nuclear loci.

[shweta shukla]

Thank you very much!

I have substitution rate in per year . Can you please explain why to take "7.1e-3" value for per year. also I have to find out divergence time can explain where I can look in Tracer software for the same.

Thank you!

[sshweta]

Should I set the rate in substitution rate block or clock.rate block????

Kindly reply as soon as possible

[Huw A. Ogilvie]

If your substitution rate is 7.1x10^-9 per year, then that's the same as 7.1x10^-3 per million years. So if you use a rate of 7.1x10^-3, that will calibrate your tree to be in units of millions of years. On a computer, "e-3" is shorthand for "x10^-3", so it's easy to type "7.1e-3" instead.

The divergence times are logged on the tree rather than tracer (unless you have an MRCA prior for the node in question). Divergence times for all nodes can be viewed in densitree, which is included with BEAST2.

The substitution rate and clock rate values are multiplied during the analysis, so 1.0 for substitution rate and 7.1e-3 for clock rate is exactly the same as 7.1e-3 for substitution rate and 1.0 for clock rate. I would suggest using substitution rate to estimate relative rates among loci (by making sure substitution rates have initial values of 1, are estimated, and "fix mean substitution rate" is set), then link the clocks and fix the clock rate to 7.1e-3. This way 7.1e-3 will be enforced as the true average substitution rate across all sites in your data.

- Huw

[sshweta]

Thanks a lot once again!

But I think my tree does not showing the divergence time  to me  even I have put all the value in BEAST2 then how could I interpret from tree??? Any other parameter should I have to use for getting divergence time??? 

There is one more confusion I have about the prior selection as I have datasets of the plant and I want to know their divergence time, which model I have to use to calculate this could you have some idea regarding this ????

[Huw A. Ogilvie]

I would recommend first learning about divergence dating from the tutorials available online. For BEAST2, you can try the "Divergence dating" and "Divergence dating with sampled ancestors" tutorials, available on the BEAST2 web site - http://www.beast2.org/tutorials/

For StarBEAST2, you can follow the tutorial available on GitHub - https://github.com/genomescale/starbeast2/releases/latest

[sshweta]

Thank you!

[sshweta]

Hi,

After following these tutorials I have concluded some results. Here I am attaching my recent result Could you see the files and let me know Is this would be the divergence time of my datasets??? or I have to do something else??

[sshweta]

where I could create TMRC in Beauti2?????