BUCKy users Google Group

Convergence and mixing

lp...@nyu.edu — Sat, 15 Jun 2013 16:10:40 UT

Hi all,
Sorry for bothering the list again but I have a problem in understanding my
results: I ran a dataset with 13 taxa and 20 loci (complete sampling = no
missing taxa/loci). When I analyzed the dataset using bucky with various
alpha values (1, 10, 100), I have CFs=1.00 on all nodes but when I set the

Re: how to treat missing taxa?

lp...@nyu.edu — Thu, 13 Jun 2013 17:39:34 UT

Hey! Thank you very much Cecile and Jenna. Your comments were very helpful.
It seems that my dataset is pretty strong even with missing data but I just
wanted to be sure it made sense.
Luca

Re: how to treat missing taxa?

jennom...@gmail.com (Jenna Morgan Lang) — Thu, 13 Jun 2013 15:44:40 UT

I've had to deal with this issue for a recent publication:

[link]

Hope this is helpful!
Jenna

Re: how to treat missing taxa?

a...@stat.wisc.edu (Cecile Ane) — Thu, 13 Jun 2013 15:25:41 UT

My experience is that missing taxa do not affect the topology of the
primary concordance tree very much. However, the number of taxa does
affect the estimated concordance factors, with more taxa leading to
lower concordance factors in general (and I trust concordance factors
from few taxa more than those from more taxa).

how to treat missing taxa?

lp...@nyu.edu — Thu, 13 Jun 2013 14:12:31 UT

Hi,
I have a quick question about missing data. I have a matrix of 27 loci with
16 taxa but some taxa are missing for several loci. At first I ran mrbayes
with the command *unlink topology* and I got *.t files* for each locus I
wanted to analyze. However, I wasn't sure if the missing taxa were
affecting my analyses so I ran mrbayes independently for each locus

Re: Problems with the website to select alpha prior

a...@stat.wisc.edu (Cecile Ane) — Thu, 14 Feb 2013 21:16:31 UT

Hi Brent,
This webform is hard to fix, each time the server OS or other software
is updated. If you are an R user, you can get the visualization yourself
using R code now provided here:
[link]
[link]

Re: Problems with the website to select alpha prior

lar...@stat.wisc.edu (Bret Larget) — Thu, 14 Feb 2013 05:18:19 UT

Actually, i habe the same problem too. I was not looking at the graph
carefully. We'll look into it.

-Bret

Re: Problems with the website to select alpha prior

lp...@nyu.edu — Wed, 13 Feb 2013 22:23:13 UT

Hey Brent,
Thank you very much for your quick reply. Unfortunately that doesn't solve
the problem for me. I tried also different computers and browsers
(including Safari) but the result is always the one I posted earlier: once
I set my parameters and I click "recalculate" I get the results
for genes=5500; taxa=17; alpha=10. I have tried many different combinations

Re: Problems with the website to select alpha prior

lar...@stat.wisc.edu (Bret Larget) — Wed, 13 Feb 2013 16:26:26 UT

This web page works just fine for me using Safari.

[link]

-Bret Larget

Problems with the website to select alpha prior

lp...@nyu.edu — Wed, 13 Feb 2013 16:01:34 UT

Hi all,
I am a new user of bucky and I'm encountering some problems in using the
interactive web tool to select alpha in my analyses. I tried to set my
number of individuals (24) and loci (23) in the system but no matter what
numbers I select in the fields the same prior distribution is displayed

Re: BUCKy runs but terminates while writing run1.concordance

yoshanmood...@googlemail.com (Yoshan) — Thu, 07 Feb 2013 18:20:27 UT

hi Cecile
has this problem been fixed in the unix version? I have exactly the same
problem as jeff, were the program hangs at the concordance file stage, then
it gives and error and stops. i edited and recompiled as you suggested, but
the problem does not go away. infact now the program comes up with a

Re: follow up outgroup Q

elaineeguev...@gmail.com (Elaine Guevara) — Sun, 09 Dec 2012 17:28:21 UT

Hi Cecile,

the CF is not listed in the bucky output file, so it may just be the way it
appears in figtree (the .5 being the assigned "branch length"). I will just
delete that value from the visualization of my concordance tree.

Thanks,
Elaine

Re: follow up outgroup Q

a...@stat.wisc.edu (Cecile Ane) — Sun, 09 Dec 2012 17:00:06 UT

Hi Elaine,
If the output is a single taxon, it defines a 'trivial' bipartition,
because bucky considers unrooted trees. It should not receive any
concordance factor. I don't know how you got CF of .5 in the first case
you listed.
Cecile.

follow up outgroup Q

elaineeguev...@gmail.com (E Guevara) — Fri, 07 Dec 2012 20:50:31 UT

Hi all,
I have another question regarding outgroups/rooting. I notice that the node
joining my outgroup to the tree in the way I have rooted it has a CF of .5.
Is this just an arbitrary number? As I rerooted the tree, this split is not
listed in the .concordance file and the CF does not have a confidence

Alpha Prior

yohan_pil...@hotmail.fr (Yohan) — Thu, 29 Nov 2012 21:26:42 UT

Hello,
The website that calculate the prior distribution on the number of distinct
trees with different value of alpha does not seem to work anymore.
In my case, it seems that the prior alpha has an important impact on the
number of clusters found by BUCKY. What is the best way to choose alpha?
I have 14 genes for 6 species. For alpha=0,1; I recover 2.25 cluster,