Hi Roger,
Fast Solution: use pathd8 or r8s or treePL. Not in R... pathd8 should be the 
easiest one since you just need a ultrametric tree... 1 sec in parthd8

Cheers, Franz




> On 21.09.2015, at 16:21, Fabian Roger <[email protected]> wrote:
> 
> hej everyone
> 
> I have a large phylogeny based on 16S RNA OTU data (> 7000 OTUs).  In order 
> to calculate some metrics of phylogenetic diversity I need to make the tree 
> ultrametric. I thought about using the chronos function from the ape package 
> in R. 
> 
> However I run into two problems: the main problem is that for the full tree 
> the function fails with the error "Error in nlminb(start.para, f, g, control 
> = opt.ctrl, lower = LOW, upper = UP) : NA/NaN gradient evaluation”
> 
> I have seen several question regarding this error on the mailing list, but I 
> haven’t seen a solution. Could anybody help with that? Not that if I run 
> small subsets of the full tree, it works without problems!
> 
> Also, I am concerned about runtime issues. A quick simulation suggests that 
> runtime roughly scales linearly with the number of OTUs and that 10 000 OTUs 
> should run in below 1 h (although I only simulated trees up to 140 OTUs…) if 
> that’s true it wouldn’t be all that bad, but does anybody know or has 
> experience on if thats actually true? 
> 
> thanks a lot for your help!
> 
> best
> 
> Fabian
> _______________________________________________
> R-sig-phylo mailing list - [email protected]
> https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
> Searchable archive at http://www.mail-archive.com/[email protected]/

_______________________________________________
R-sig-phylo mailing list - [email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/[email protected]/

Reply via email to