Re: [Dart-help] help with windowlicker

[marcin j. <mar...@gm...>]

Excellent, its working now!

I'll have the tree data soon and may try a run with as well, but best
if I could know that the other tree is different from the xrate one
with the Jukes-Cantor model. Can I get the Jukes-Cantor model
tree/distances from xrate?

marcin

On Thu, Apr 24, 2008 at 6:51 PM, Ian Holmes <ih...@be...> wrote:
> marcin joachimiak wrote:
>  > Thanks!
>  > Initially the only environment changes I made was adding the perl
>  > module dir to PERLLIB.
>  >
>  > I've now created the DARTDIR variable and there is some progress but
>  > now I get a different error message:
>  >
>  > -bash-3.00$  perl /usr2/people/gtl/bin/dart/perl/windowlicker.pl
>  > DvH_DP4_G20_mVISTA.stock
>  > [stderr] ERROR: Every input alignment needs to be annotated with a
>  > phylogenetic tree.
>  > [stderr] Given that this isn't the case, I need a rate matrix in order
>  > to do some ad-hoc phylogeny
>  > [stderr]  (neighbor-joining followed by EM on the branch lengths, if
>  > you must know.)
>  > [stderr] I want to get that rate matrix from the tree estimation grammar file.
>  > [stderr] However, I haven't been told about any such file, so I'm bailing.
>  > xrate returned with nonzero exit status (invoked using:
>  > /usr2/people/gtl/bin/dart//bin/xrate )
>
>  Do you already have a phylogenetic tree for these sequences? If so, and
>  you can write the tree in New Hampshire format, then you can put it
>  directly into your Stockholm alignment file, like so:
>
>  # STOCKHOLM 1.0
>  #=GF NH ((seq1:0.01,seq2:0.03):0.04,seq3:0.05);
>  seq1 AAGGGCCC
>  seq2 AGGGGCCT
>  seq3 GAGGGCCT
>  //
>
>
>  If not, you need to specify a grammar file for tree estimation. This is
>  not the same as the grammar file for alignment annotation (confusing, I
>  know). Currently tree estimation can only be done with simpler grammars
>  than the RNA gene prediction model we are using.
>
>  We typically use the Jukes-Cantor model for tree estimation. This is
>  encoded as a grammar file in "$DARTDIR/grammars/jukescantor.eg". To
>  specify the tree-estimation grammar, you use the "-e" option (the "-g"
>  option is for the alignment annotation grammar).
>
>  So, what you want is probably something like this:
>
>
>  perl /usr2/people/gtl/bin/dart/perl/windowlicker.pl
>  DvH_DP4_G20_mVISTA.stock -- -e $DARTDIR/grammars/jukescantor.eg
>
>
>  hope this helps! if not, send along the alignment file and I'll take a look.
>
>  Ian
>
>
>
>
>
>
>  >
>  > I've also tried specifying the grammer file for xrate but apparently
>  > unsuccessfully:
>  >
>  > -bash-3.00$  perl /usr2/people/gtl/bin/dart/perl/windowlicker.pl
>  > DvH_DP4_G20_mVISTA.stock -- -g ncRna_v22.eg
>  > [stderr] ERROR: Every input alignment needs to be annotated with a
>  > phylogenetic tree.
>  > [stderr] Given that this isn't the case, I need a rate matrix in order
>  > to do some ad-hoc phylogeny
>  > [stderr]  (neighbor-joining followed by EM on the branch lengths, if
>  > you must know.)
>  > [stderr] I want to get that rate matrix from the tree estimation grammar file.
>  > [stderr] However, I haven't been told about any such file, so I'm bailing.
>  > xrate returned with nonzero exit status (invoked using:
>  > /usr2/people/gtl/bin/dart//bin/xrate -g ncRna_v22.eg)
>  >
>  >
>  > marcin
>  >
>  >
>  >
>  > On Thu, Apr 24, 2008 at 5:07 PM, Ian Holmes <ih...@be...> wrote:
>  >> Hi Marcin,
>  >>
>  >>  first, sorry I didn't reply to your earlier email to me, been a bit busy
>  >>  here. There is actually a dart-help mailing list...
>  >>
>  >>  http://biowiki.org/DartMailingLists
>  >>
>  >>  ...but it'll pretty much end up in the same inboxes if you email
>  >>  holmes-lab instead ;)
>  >>
>  >>
>  >>  second, the error you're getting means that xrate couldn't be run. If
>  >>  you just type "xrate -h" from your command prompt, do you get a help
>  >>  message? Or do you have to type in the entire path to xrate, e.g.
>  >>  "/nfs/src/dart/bin/xrate -h" or whatever?
>  >>
>  >>  If the latter, then it's likely that windowlicker just can't find the
>  >>  xrate binary. You can test this by typing "windowlicker.pl -h" and
>  >>  looking at the "-x" option, whose default value is the path that
>  >>  windowlicker is trying to use.
>  >>
>  >>  Best fix for this is to set the DARTDIR environment variable to the root
>  >>  of your DART installation (/nfs/src/dart in the above example). You can
>  >>  also use windowlicker's -x option to tell it the path to the xrate
>  >>  executable.
>  >>
>  >>
>  >>  third, if the above doesn't fix your problem, can you email me the
>  >>  actual file you're having problems with (if it's huge, then gzip or bzip
>  >>  it first) so I can try to reproduce the issue.
>  >>
>  >>
>  >>  Hope this helps!
>  >>  Ian
>  >>
>  >>
>  >>
>  >>  PS I am still trying to get around to reading your guide tree
>  >>  manuscript. It will happen though!
>  >>
>  >>
>  >>
>  >>
>  >>  marcin joachimiak wrote:
>  >>  > hello,
>  >>  >
>  >>  > I have the DART package built and running at the syntax help level.
>  >>  > Formatting the FASTA file to Stockholm was successful.
>  >>  > I've tried running windowlicker on the stocholm data and got:
>  >>  >
>  >>  > -bash-3.00$ perl /usr2/people/gtl/bin/dart/perl/windowlicker.pl
>  >>  > DvH_DP4_G20_mVISTA.stock
>  >>  > xrate failed (check command-line options? they were: xrate )
>  >>  >
>  >>  > So I assume there are more required arguments? And how would I supply
>  >>  > the reference grammars?
>  >>  >
>  >>  > thanks,
>  >>  > marcin
>  >>
>