FDPPDIV-fossil calibration&more

[sah...@iisertvm.ac.in]

Dear Tracy,

As a new user of the FDPPDIV, I have few questions to run the analysis for which I did not find any reliable answer yet.

Since FBD assumes the occurrence of fossil at a single time point, it is appropriate to sample a time point from the uniform distribution of the age ranges of the concerned fossil. This is what you have mentioned in your PNAS 2015 paper. But in a recent paper by Gavryushkina etal 2015 in arXiv.org, if I am right, the group have used the age ranges rather than a single time point. If so, then how to assign the prior for each fossil age calibration?

As I found, in most of the FBD analysis, a command "-tga" have been used. What does it stand for?

By default the -clock considers strict clock as tree prior. How to change it to uncorrelated relaxed clock?

How to assign FBD model to the analysis? As I found the -npr command assigns either uniform, yule, cbd or cbd fix, but no" fbd" option is available.

As the topology is predefined, we do not have to feed the partition file and site model to the FDPPDIV analysis. Am I right?

Would you please suggest any manual for FDPPDIV, if available?

Thanks and Regards,

Ranjit Kumar Sahoo

[Ranjit Kumar Sahoo]

One more question: In node dating analysis, calibrated node(s) were assigned with the node age and prior density distribution. But in FBD analysis, though we do not assign the age constraints to the nodes, the fossil calibration file mentions the specific node for each fossil input. Then how precise the node assignment to a given fossil is necessary for the precise age estimation in FBD analyses? 

e.g.

Let's I have a single fossil with known age to calibrate the tree ((T1, T2), (T3, (T4, T5))) 

Due to incomplete information from the fossil I am uncertain to assign the fossil to the node (T4, T5). But I am sure that the fossil is one of the member of the group containing T3, T4, T5. Thus assigned the fossil calibration as 

-t   T3    T5    50

Does such precise assignment of nodes to each fossil is necessary for FBD analysis? What about the fossils for which we have incomplete taxonomic information?

Thanks,

[Tracy Heath]

Hi Ranjit,

Below are answers to your questions:

1. To use age ranges, you either have to download the development code for DPPDiv, or use BEAST2. At the moment, I recommend BEAST2.

2. "-tga" is simply a flag to indicate the FBD model. Use this instead of the "-npr" arguments to set the FBD model.

3. By default DPPDive assumes a DPP prior on branch rates, you can set an uncorrelated relaxed clock by using the "-urg" flag instead of "-clok".

4. I'm not sure what you mean by "partition file".

5. There is no manual for DPPDiv.

Cheers,

Tracy

--
You received this message because you are subscribed to the Google Groups "dppdiv-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dppdiv-users...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Tracy Heath]

Hi Ranjit,

For any analysis using fossils, it is ideal to assign them as precisely as possible. You can account for uncertainty in their assignment by indicating they are descended from a deeper node. The more uncertainty in their placement, the greater chance that this will affect your node age estimates. In the example you gave above, whether you are using the fossil in an FBD model or as a node calibration, if you are certain the fossil is a descendant of the most-recent-common ancestor of species T3 and T5, then this is the node you would specify. If you are uncertain, then a deeper node is appropriate.

Cheers,

Tracy