tag:blogger.com,1999:blog-1871542942842750523.post8543197314948032508..comments2023-07-24T10:40:57.739-04:00Comments on dechronization: R Tip: Labeling Trees w/ Posterior Probability and Bootstrap SupportGlorhttp://www.blogger.com/profile/17707197225963721646noreply@blogger.comBlogger9125tag:blogger.com,1999:blog-1871542942842750523.post-26780717767938637992014-09-11T10:15:58.932-04:002014-09-11T10:15:58.932-04:00use Mega6 !use Mega6 !Anonymoushttps://www.blogger.com/profile/14762535124122854493noreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-2181363450275410852013-08-30T09:39:32.574-04:002013-08-30T09:39:32.574-04:00Hi Rich-- does this function still work for you, o...Hi Rich-- does this function still work for you, or have you since updated it? I just tried on R v3.0.1 and am getting a series of errors. Seems like a useful tool so I'd like to give it a try!<br />ThanksNate Uhttps://www.blogger.com/profile/07604998134000908594noreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-50479444483096136322008-11-03T16:57:00.000-05:002008-11-03T16:57:00.000-05:00Although I have yet to try this approach, Casey Du...Although I have yet to try this approach, Casey Dunn suggested that <A HREF="http://code.google.com/p/phyutility/" REL="nofollow">Phyutility</A> - the application he wrote with Stephen Smith - will also perform the type of analyses implemented by Jeet's SumTrees programGlorhttps://www.blogger.com/profile/17707197225963721646noreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-79358299130463465642008-10-18T23:37:00.000-04:002008-10-18T23:37:00.000-04:00Rich:Yep, in fact the original motivation was to d...Rich:<BR/><BR/>Yep, in fact the original motivation was to do exactly what Todd requested. But then I later decided that I sometimes might be interested in summing clade posterior probabilities over multiple tree files as well, but *not* mapping it onto an ML tree (i.e., I wanted clade support but wanted to integrate out topology). So I added the burn-in and consensus tree option. When I get around to the next version release, I'd like to add an option to summarize the posterior probabilities of node ages (=tip-to-node distance) as opposed to just edge lengths as it does now.<BR/><BR/>By the way, great site---it's definitely one of my regular/daily ports-of-call now!<BR/><BR/>-- jeetAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-53285165642249877672008-10-18T22:18:00.000-04:002008-10-18T22:18:00.000-04:00Jeet - Thanks for pointing out your program. I ha...Jeet - Thanks for pointing out your program. I have yet to give it a try, but it sounds very useful. It will be nice to able to plot from the original data rather than from the consensus trees, as I've done with the R script. It sounds like it will also make it easier to do what Todd has suggested and plot support values on the tree with the highest ML score.Glorhttps://www.blogger.com/profile/17707197225963721646noreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-10537314858917513882008-10-18T20:50:00.000-04:002008-10-18T20:50:00.000-04:00Hi guys,I wrote SumTrees specifically for this pur...Hi guys,<BR/><BR/>I wrote SumTrees specifically for this purpose.<BR/><BR/>http://jeetworks.org/programs/sumtrees<BR/><BR/>Takes one or more tree files as input, and a target tree file, and plots split support based on the former onto to the latter.<BR/><BR/>It also has some other options such as skipping a burn-in, composing a consensus tree if a target tree is not given, etc.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-9475761249649322732008-10-18T20:41:00.000-04:002008-10-18T20:41:00.000-04:00Have you tried FigTree? It's quite clever as i...Have you tried <A HREF="http://tree.bio.ed.ac.uk/software/figtree/" REL="nofollow">FigTree</A>? It's quite clever as it'll let you plot anything stored as a comment in the Newick description on the tree nodes/edges. So, you give it something like this:<BR/><BR/>(taxa1[bootstrap=100&pp=0.99],taxa2[bootstrap=85&pp=0.88])...etc<BR/><BR/>..thenyou can choose what/how to label the nodes. BEAST automatically stuffs every single parameter into the tree description for FigTree to use, but it's fairly trivial to reverse engineer other software to do the same.<BR/><BR/><BR/>--SimonAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-20967104016305600932008-10-18T15:53:00.000-04:002008-10-18T15:53:00.000-04:00There is an r function in ape, ladderize(), that w...There is an r function in ape, ladderize(), that will ladderize a tree for plotting. If you run your tree through that before running Rich's script, I think it will do what you want.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1871542942842750523.post-7855731085401691952008-10-18T15:03:00.000-04:002008-10-18T15:03:00.000-04:00Thanks, Rich - this (you giving detailed instructi...Thanks, Rich - this (you giving detailed instructions) is a great way to slowly learn R without having to read the giant R book much.<BR/>Can you rotate branches or ladderize?<BR/>This is important for showing multiple similar trees or for getting the trees to match geography.<BR/>Also, if the Bayes consensus has polytomies, I would prefer showing the Bayes tree with best ML score, but that can be done easily.Todd Jackmanhttps://www.blogger.com/profile/02075021843436668482noreply@blogger.com