Dear EvolDir members, I'm studying the phylogeography of a sea urchin species by sequencing the entire mitochondrial gene Cyt b. Assignment analysis with BAPS revealed the presence of three sympatric clusters with maximum statistical support (P=1). To make some considerations about the origin and spread of the clusters I'd like to estimate the divergence time among the three clusters. I would greatly appreciate if someone could suggest me the way and the software to proceed in this estimation? Answers: - Hi Michele, would suggest MDiv. Please could I have the summary of replies? thanks and good luck, Ute - use BEAST (rambaut & drummond et al) Mark Schultz Charles Darwin University C/- Arafura Timor Research Facility PO Box 41775 Casuarina NT 0811 Australia Ph: +61 (0)8 8920 9292 Fax: +61 (0)8 8920 9222 - Ciao Michele, ho letto il tuo post su evoldir; io lavoro a ferrara e mi sono (e talvolta ancora) occupato di tempi di divergenza, barcodig e cose del genere; di solito quando faccio datazioni molecolari uso Beast (gli autori sono Drummond e Rambaut beast.bio.ed.ac.uk/Main_Page); altrimenti puoi provare con Mrbayes (anche se è più complicato gestire l'output) oppure con multidivtime (del gruppo di Thorne http://statgen.ncsu.edu/thorne/multidivtime.html) o r8s (di Sanderson http://ginger.ucdavis.edu/r8s/)) ma usando questi ultimi due approcci devi avere già un albero sensato che rappresenti le relazioni tra le tue sequenze e in generale implementano relaxed clock models (mentre con mrbayes e beast puo mettere l'orologio "strict"). Oppure puoi semplicemente calcolare le distanze tra gruppi con Mega e avere una idea di massima della situazione. Spero di esserti stato utile e se hai bisogno di altre informazioni scrivimi pure. Ciao, Stefano - Hi Michele, Without knowing much about your data, I'd suggest IMa (by Nielsen and Hey) and/or BEAST (Drummond et al). Are the clulsters very different? Just by stochastic processes you could have the appearance of clades while the population is actually panmictic. In my thesis work I used MDIV, which is like IMa but simpler, and BEAST to look at divergences among frogs. see http://www.biomedcentral.com/1471-2148/7/247/abstract good luck, Kathryn ------------------------------ Kathryn R. Elmer, Ph.D. Humboldt Postdoctoral Fellow Evolutionary Biology (LS Meyer) Dept. of Biology University of Konstanz D-78457 Konstanz Germany phone: +49 (0) 75 31 88 37 10 (from abroad, no (0)) fax: +49 (0) 75 31 88 30 18 e-mail: kathryn.elmer@uni-konstanz.de website: http://www.evolutionsbiologie.uni-konstanz.de/~kathryn/ - You really need to have an external marker, like a biogeographic event. MtDNA evolution rates can vary quite a lot and can be very different between relatively closely related lineages. Certainly this is true in insects, which I work in; perhaps you should look into whether urchins might be more consistent. However, I would certainly not assume that a rate obtained for a distant group will apply to yours. Karl ====================Karl Magnacca Postdoctoral Researcher Department of Zoology Trinity College, Dublin 2 Ireland - ciao Michele riguardo alla tua richiesta su evoldir i programmi piu' usati per divergence time siano MrBayes e Beast. mrbayes e' facile da usare, mentre Beast fornisce molte piu' opzione, ma e' piu' difficle da usare. consiglio anche Tracer per analizzare gli output di entrambi i programmi. pero' se i tuoi cluster sono interni alla specie (cioe' stiamo parlando della stessa specie) allora un approccio cladistico e' impossibile perche' ce' orizzontal transfer. In quel caso, a detta di un mio collega che studia divergence time, non ce' ninete da fare. inoltre lui sottolinea che avresti cmq bisogno di un grosso numero di sostituizioni nucleotidiche, diciamo un centinaio per ramo, ma dubito che cytb per organismi cosi correlati proti un segnale di questo tipo. magari ce' ma qui non lo sappiamo. ciao e in bocca al lupo. - Hi Michele, The program r8s (Sanderson, 2002 MBE 19:101; Sanderson, 2003, Bioinformatics 19:301) will allow you to do this based on adding another sequence into your phylogeny for which the divergence time is known. You then constrain the node that the div time is known for (fossil records are the best source of this sort of data) and the remainder of the node ages are calculated for you Good luck --mark chapman -------------------------------- Mark A. Chapman mchapman@plantbio.uga.edu My webpage: http://markachapman.googlepages.com/ The lab webpage: www.theburkelab.org -------------------------------- University of Georgia Department of Plant Biology Miller Plant Sciences Bldg. Athens, GA 30602 -------------------------------- -Hola Michelle, You could use the software BEAST which develops Bayesian inference, estimating the topology and divergence time simultaneously. http://beast.bio.ed.ac.uk There you can check a more formal description of this software. The developers provide a very useful manual. It has an easy interface (mac linux or windows). As soon as you became familiarized with BEAST you will need fossil estimates or some estimate of the substitution rate in order to get good estimates. Let me know if it works or if you find it useful. I could give you more details about this particular soft. There are other applications that can provide molecular dating but I found them "compromising" You could search for the better substitution model that fits your data here; http://distributed.cs.nuim.ie/multiphyl.php. Ciao, Andrés Sección Evolución, Facultad de Ciencias Uruguay ---------- Michele Barbieri Ph.D. student Università di Pisa Dipartimento di Biologia Unità di Biologia Marina ed Ecologia Via A. Volta 6, 56126, Pisa - Italy (I) tel: +39 050 2211407 fax: +39 050 2211410 mbarbieri@biologia.unipi.it Michele Barbieri