Ns in RAxML

[Emily Gillespie]

I've used RAxML a fair amount, but something unusual happened this time. I use Geneious for alignment, and export as .phy file. I do also use a partition file. None of this has ever been a problem. This time, however, I got a 'cannot read alignment file' error. I've checked to make sure all the underscores are right in the taxon names, etc... so I thought maybe RAxML doesn't like end gaps. So I exported as fasta (both with gaps as Ns and gaps as - ) and reimported, then exported again, this time as .phy. The one with gaps as Ns was accepted, but not the one with gaps as - . Question is, will Ns impact how the topology is determined, or just branch lengths? I could see it going either way. Eventually, I need branch lengths, but for now, topology is ok. It's preliminary.

Thanks for any insight.

[Alexandros Stamatakis]

Dear Emily,

Ns and gaps are treat mathematically in exactly the same way in RAxML and most other likelihood based codes
(please use the search function to find previous post regarding this issue).

As to why RAxML can't parse the file I'd need the input alignment and the RAxML version you are using.

All the best,

Alexis

--
Dr. Alexandros Stamatakis
Research Group Leader HITS, Heidelberg
Adjunct Professor, Dept. of Ecology and Evolutionary Biology, University of Arizona at Tucson
www.exelixis-lab.org

[Emily Gillespie]

Hi Alexis,

I had searched the google group and the one discussion I'd found didn't seem to answer my question exactly. Maybe I didn't realize that it did.

I've never had any trouble at all getting files to load and run. I'd be grateful if you could look at my files. 

I've attach three; one that is a just 'plain', one that ends in 'dash' and one that ends in 'N'. They are exported differently. The 'plain' one is just exported as usual from Geneious as a .phy file. When it wouldn't load, I then exported it as a fastA with gaps as dashes, then reimported and re-exported as .phy. It also failed, so the third one, I exported as fastA with gaps as Ns, reimported and then re-exported as .phy. 

The 'plain' and 'dash' simply give me a 'raxml could not read the alignment file' error. The 'N' one will load, but I very quickly get a 'raxml has finished' email, and the link indicates a bootstrap file and job file, but both are empty.

I've also attached my data partition file, in case it helps.

I was using the raxml on blackbox.

I very much appreciate any guidance. I'm totally at a loss.

--
Emily Gillespie, Ph.D.

Research Associate
Department of Biology
Wake Forest University
Winston-Salem, NC

gill...@wfu.edu

[Fidel Botero]

Hi Emily,

I explored your plain file and RAxML send me the message "Bad base (_) at site 1 of sequence 511".
Looking into the file, there is in fact a space in "Cavendishia_capitulata@_ACC". Same thing occurs for sequence 538 "Diogenesia_racemosa@_ACC".

RAxML takes the first space it founds as the separator between taxa names and sequence first position. Thus, it was finding the first base for these taxa to be a "_" which gives an error. I corrected it (names should look as  "Cavendishia_capitulata_ACC" and "Diogenesia_racemosa_ACC" respectively) and it seems to run properly.

Hoping this will help you.

Fidel 

2012/7/2 Emily Gillespie <emilylg...@gmail.com>

--
Fidel Botero
Biologiste
Doctorant Université Montpellier 2
Équipe Phylogénie et Évolution Moléculaires
Institut des Sciences de l'Évolution de Montpellier - ISE-M (UMR 5554 UM2-CNRS)
04 67 14 92 00

[Emily Gillespie]

Hi Fidel,

THANK YOU so much. I totally could not see those spaces. Even after you pointed them out, I couldn't see that they were there at first. I was so frustrated!

Thank you again. I suspect it will run now.

Emily

[Fidel Botero]

Just in complement :

For the Ns file, it runs because the space has been filled with another "_".

For the dashed file, the space is also there.

After correcting, both files seem to work correctly.

This should also explain why your partition file is not working properly as positions were not being properly read.

Fidel

2012/7/2 Fidel Botero <fidel....@gmail.com>

[Fidel Botero]

You're welcome ! ;)

[Alexandros Stamatakis]

Why don't you download RAxML (the latest version from github) and run the -f c option that checks
if the alignment format can be parsed.

If you then still get an error, please send me the alignment files again and I will try to figure out
what the problem is.

Alexis

On 07/02/2012 06:55 PM, Emily Gillespie wrote:
> Hi again Fidel,
>
> Now, the problem is that I get back a result notice almost immediately via
> email, but there are no files. I have never had these problems in the past.
> It's been a while, maybe I am forgetting something stupid.
>
> I have uploaded the .phy file with the corrections. I have left the CAT
> model in place. I have not checked the protein box. I have identified an
> outgroup. I have not included a constraint or backbone file. I have
> included my dna partition file (I doublechecked to see that it is formatted
> according to the manual--which is different from the example on the
> blackbox website). I have checked the maximum likelihood search box, which
> makes the invariable sites box disappear.
>
> What am I missing?

>
> Emily
>
> On Mon, Jul 2, 2012 at 12:13 PM, Fidel Botero <fidel....@gmail.com>wrote:
>

>> Doctorant Universit� Montpellier 2
>> �quipe Phylog�nie et �volution Mol�culaires
>> Institut des Sciences de l'�volution de Montpellier - ISE-M (UMR 5554

>> UM2-CNRS)
>> 04 67 14 92 00
>>
>
>
>

[Emily Gillespie]

I'll do that. It would be good to learn how to use the stand-alone version anyway.

Thanks,

eg

Doctorant Université Montpellier 2
Équipe Phylogénie et Évolution Moléculaires

Institut des Sciences de l'Évolution de Montpellier - ISE-M (UMR 5554

UM2-CNRS)
04 67 14 92 00

--
Dr. Alexandros Stamatakis
Research Group Leader HITS, Heidelberg
Adjunct Professor, Dept. of Ecology and Evolutionary Biology, University of Arizona at Tucson
www.exelixis-lab.org

[Alexandros Stamatakis]

There's an on-line tutorial at:

http://sco.h-its.org/exelixis/hands-On.html

Alexis

>>>>> alexandros.stamatakis@gmail.**com <alexandros...@gmail.com>>

>>>> Doctorant Universit� Montpellier 2
>>>> �quipe Phylog�nie et �volution Mol�culaires

>>>> Institut des Sciences de l'�volution de Montpellier - ISE-M (UMR 5554

>>>> UM2-CNRS)
>>>> 04 67 14 92 00
>>>>
>>>>
>>>
>>>
>>>
>> --
>> Dr. Alexandros Stamatakis
>> Research Group Leader HITS, Heidelberg
>> Adjunct Professor, Dept. of Ecology and Evolutionary Biology, University
>> of Arizona at Tucson
>> www.exelixis-lab.org
>>
>>
>>
>
>

--

[Fidel Botero]

Hi again,
I've tried directly through the Blackbox interface and like you, it didn't work.
I then tried through the CIPRES mirror, and it's running without error messages... Could you try like this ?

Hope you can solve it !

[Emily Gillespie]

Hi,

Silly question I guess, but which of the many output files I got from the Cipres server has my tree with bootstrap values? It never has run properly through BlackBox, but may have run at Cipres. 

Many thanks.

[Fidel Botero]

Hi Emily, 

I also tried in the CIPRES portal, it ran and then I got some files with an error output among them. Only after correcting the things I mentioned before (mainly the partition file) it ran properly. 

I guess you did it as a 'guest', without registering... :p I also did it and there's no way of recovering the work after. The best is to register and run it from your account.

As it runs in a paralelized way, you should obtain several files named RAxML.bootrsap_result.PID.# corresponding to the different nodes, and one named simply RAxML_bootstrap.result. This one contains the 100 trees of the bootstrap sampling. Otherwise, I can't find the best tree on which bootstraps should be mapped... I'm doing a new try to see if it generates it...

Fidel

2012/7/4 Emily Gillespie <emilylg...@gmail.com>

[Alexandros Stamatakis]

It should be called RAxML_bipartitions or something similar,

Alexis

>>>>>>> alexandros.stamatakis@gmail.**com <alexandros...@gmail.com>>

[Emily Gillespie]

Hi Alexis and Fidel, 

I thought that was the name, too (the bipartition file). At least that's what it had been when I ran it through black box. No such file seems to exist in the Cipres output files.

I was actually running as a registered user at Cipres. To be honest, I can't figure out what the tree file is in ANY of the applications I've tried there (PAUPRat, GARLI, BEAST...). I'm whatever the opposite of a programmer is, so I suspect I just don't recognize it, but I literally tried importing EVERY output file (even the ones I was sure was wrong) into FigTree and the best I got was an invisible non-tree with a scale bar. 

I'm working on cutting down my dataset now, hoping it will run faster, so it will at least not take as long to figure out if it ran properly. (And I'm running out of time to do multiple runs). I may ask you guys to take a look if it doesn't run properly now.

Any idea why the Blackbox output link (email) now sends me to that list of all of the current jobs, instead of to the page with just several of my files? Even if my file runs next time, I can't retrieve it.

Thank you for your continued help.

Emily

alexandros.stamatakis@gmail.**com <alexandros.stamatakis@gmail.com>>

[Mark Miller]

Hi Emily,

 

Thanks for reporting that it is hard to figure out which files have trees in CIPRES. I will look in to figuring out how to make that work better.

Also, I am happy to help if you contact me through the CIPRES user list.

 

If you didnt see a bipartions file, there is definitiely n issue with the run, and if you diudnt see a best.tree file, there is probably an issue

also. If you want to contact me directly, or through CIPRES issue reporting, I would be happy to help you sort it out.

 

Best,

Mark

 
On Wednesday, July 4, 2012 4:41:33 AM UTC-7, Emily Gillespie wrote:

Hi Alexis and Fidel, 

I thought that was the name, too (the bipartition file). At least that's what it had been when I ran it through black box. No such file seems to exist in the Cipres output files.

I was actually running as a registered user at Cipres. To be honest, I can't figure out what the tree file is in ANY of the applications I've tried there (PAUPRat, GARLI, BEAST...). I'm whatever the opposite of a programmer is, so I suspect I just don't recognize it, but I literally tried importing EVERY output file (even the ones I was sure was wrong) into FigTree and the best I got was an invisible non-tree with a scale bar. 

I'm working on cutting down my dataset now, hoping it will run faster, so it will at least not take as long to figure out if it ran properly. (And I'm running out of time to do multiple runs). I may ask you guys to take a look if it doesn't run properly now.

Any idea why the Blackbox output link (email) now sends me to that list of all of the current jobs, instead of to the page with just several of my files? Even if my file runs next time, I can't retrieve it.

Thank you for your continued help.

Emily

[Fidel Botero]

Hi Mark, 

I tried to help Emily by checking her files and trying to run it on CIPRES. I didn't find neither a 'best tree' file nor a 'bipartition' file. The only output files are the partial bootstraps for each paralel node and a compilation of this files in a 'bootstrap.result' file.

Hope this will give you some light...

Best,

Fidel

2012/7/6 Mark Miller <mark.ala...@gmail.com>

[Mark Miller]

Is there any chance one of you can send me the job handle for the job that failed?

Here is how to find it:

http://www.phylo.org/index.php/help/report_failed_job

 

Mark

[Elinia Lopes]

hi Mark, I found this topic and

I am having the same problems with my partition file. When add the partition no find a 'best tree' file and 'bipartition' file. I am using the Cipres portal. I selected the partition file in "use a mixed/partitioned  model? (-q)

I do no if I am correct

thank

Elinia

Em sexta-feira, 6 de julho de 2012 12h00min23s UTC-3, Mark Miller escreveu:

Hi Emily,

 

Thanks for reporting that it is hard to figure out which files have trees in CIPRES. I will look in to figuring out how to make that work better.

Also, I am happy to help if you contact me through the CIPRES user list.

 

If you didnt see a bipartions file, there is definitiely n issue with the run, and if you diudnt see a best.tree file, there is probably an issue

also. If you want to contact me directly, or through CIPRES issue reporting, I would be happy to help you sort it out.

 

Best,

Mark

 
On Wednesday, July 4, 2012 4:41:33 AM UTC-7, Emily Gillespie wrote:

Hi Alexis and Fidel, 

I thought that was the name, too (the bipartition file). At least that's what it had been when I ran it through black box. No such file seems to exist in the Cipres output files.

I was actually running as a registered user at Cipres. To be honest, I can't figure out what the tree file is in ANY of the applications I've tried there (PAUPRat, GARLI, BEAST...). I'm whatever the opposite of a programmer is, so I suspect I just don't recognize it, but I literally tried importing EVERY output file (even the ones I was sure was wrong) into FigTree and the best I got was an invisible non-tree with a scale bar. 

I'm working on cutting down my dataset now, hoping it will run faster, so it will at least not take as long to figure out if it ran properly. (And I'm running out of time to do multiple runs). I may ask you guys to take a look if it doesn't run properly now.

Any idea why the Blackbox output link (email) now sends me to that list of all of the current jobs, instead of to the page with just several of my files? Even if my file runs next time, I can't retrieve it.

Thank you for your continued help.

Emily

alexandros.stamatakis@gmail.**com <alexandros...@gmail.com>>