You can use MultiDivTime or r8s, despite they look a little bit "old"...

But the useful question could be "why don't you want to use BEAST ?" ;-)

Reltime

Dear Felix

try with r8s. Sorry, I can just say that because I haven't used yet. But there are many papers citing this software.

http://loco.biosci.arizona.edu/r8s/

I would suggest r8S as well.

r8s, although its a little old. Multidivtime and I think MrBayes can calculate divergence times

What about MrBayes? http://mrbayes.sourceforge.net/

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3329765/

Does it bother that is also bayesian method?

Treefinder (www.treefinder.de) has implemented some of the software mentioned above. Additionally it has a user friendly interface that makes the experience much easier.

best

juan

I think that Mega5 is available.

I have never used it, but in PhyML there is a kind of extra-program called Phytime that does it. To use it you have to compile PhyML with some more tags. Here you can find a guide on PhyML.

https://www.stat.auckland.ac.nz/~guindon/phytime/doc/phyml_manual.pdf

I don't know if you have explore the BEAUTI program associate with BEAST. In BEAUTI you can set all the parameters necessary to do the independent runs in BEAST. I think BEAUTI makes BEAST easier than before, and they provide in a wiki a lot of manuals for different BEAST approaches. Good luck!

http://beast.bio.ed.ac.uk/Tutorials

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3408070/

You should use BEAST. There's been lit comparing the programs and BEAST usually beats out things like R8s. My advice would be first just load your data as a single partition and run it. If that works add your calibrations or dating. If that works add your partitions, and if that works then add your real models. I slowly worked my way up when I taught myself to use it.

This paper compares R8s and BEAST

Mulcahy, D. G., B. P. Noonan, T. Moss, T. M. Townsend, T. W. Reeder, J. W. Sites, Jr., and J. J. Wiens 2012. Estimating Divergence Dates and Evaluating Dating Methods using Phylogenomic and Mitochondrial Data in Squamate Reptiles. Molecular Phylogenetics and Evolution 65(3): 974–991.

The Bayesian stuff just works better for me especially if I have some missing data or a dataset that get's stuck in a local maxima. (i.e. I change the temperature in Mr. Bayes). I don't get different topology, but sometimes better support for clades.

Also, I try to use multiple ingroup and outgroup fossil calibrations. If you are going to add calibrations and not use a molecular rate this paper is also a good read. If I grabed the right one/remember right they play around with calibrations in the following paper on Southern Beeches.

http://sysbio.oxfordjournals.org/content/early/2012/01/21/sysbio.syr116.full

you can try the molecular clock tool of MEGA 5

As soon as Beast don't solve problems for dates from 3 to 6 millions years (the level of error is huge !!!!) better to use as mentioned by some friends here PAML or PhyML which are...less friendly....but much more safe for dating. I heard that some new packages on "R" will be available soon.

Are they really better though? Just because they give a smaller range of dates does not mean it's necessarily more accurate.

I would use Beast coupled with Beauti. It is very easy to use.

Hey check out BEAST 2.0 It's much easier to set up a file in Beauti 2.0 as well with models and partitions. I didn't have to go into the XML code.

It's also much faster for me.

Surely it's best to use more than one method - preferably based on completely different concepts/assumptions, then compare the results. If they are same/similar then it's likely the results reflect the biology not the computation, whereas if they are not similar the other conclusion is possible. However the acid test is whether the results make biological sense.

BEAST is the most user-friendly software I have ever seen. Why are you looking for other one?

The RelTIME algorithm implemented in the new MEGA 6.0 (Tamura et al. 2013 doi:10.1093/molbev/mst197) seems to be just what you need. It is quick and easy to implement. It is a relative rates method for producing timetrees, so you need a calibration point or rate to post-hoc convert divergence estimates to absolute dates.

Thanks of all answers, they are very useful, in case of I need to use RelTime in MEGA6, according to the bufo arabicus.

I want to use it to calibrate my individuals divergence time according the molecular rate. In Mega

what I have to fill in these spaces in the RelTime the Taxon A, Taxon B, MRCA node, Min divergence, Max divergence.

Have I use well known calibrated sequence during the construction of the tree with my samples or can I use only my samples.

please i dont know what is mean by post-hoc convert divergence estimates to absolute dates from Michael Cunningham

Hello Hani, I've used as calibration points dates of fossil record of sequences in the outgroup in my analysis. Then taxon A, taxon B... are these sequences and the time Min. and Max is an estimate taking the date of these fossils as the Max divergence time since it was the most ancient record of each taxon.

I don't know if it is completely correct but the results are consistent with other similar works on my study group.

thanks, for bufo arabicus from Arabian peninsula I have sequenced 888bp of d loop.

in your opinion which best fossil record i can use and what are other data.

as i tried to find out some data concerning this issue but i did not find. i hope to help me in this point as my publication is waiting only for this point

Do you have any fossil record of your ingroup (Bufo arabicus) or any sequence used as outgroup?

MEGA6 probably has the best interface for this at present.

Thanks Yoannis

I dont have any records, that is my problem, so, can i use xenobus as record

or not. I dont know

Hi, guys. I am working with RelTime these days. But the result report negtive value about CI_Lower divtime. What's wrong with my settings? Is my fossils calibrations inaccurate?

I prefer Beauti as it is accurate to estimate the divergence time but it is very complicated and hard, i tried alot of times but i couldnot

So, use MEGA.6 but it is not almost accurate and its time tree appear intercalated for all samples