[EMBOSS] Bootstrap with PHYLIP

Peter Rice pmr at ebi.ac.uk
Fri Oct 21 12:21:00 UTC 2005


Dear Peter,

Peter.Robinson at t-online.de wrote:
> Dear EMBOSSers,
> 
> I would like to use the EMBASSY programs of PHYLIP such as fdnadist to do 
> bootstrap analysis etc run by a sheel script. In the menudriven version of 
> PHYLIP, many of the programs have a menu item "M" where one can tell the program 
> how many datasets to use (e.g., dnadist). In the EMBASSY version, at least some 
> of the programs do not seem to have this; it is not documented on the website, 
> and an attempt to go
> $ fdnadist -M 1000 -sequence test.fseqboot
> does not work (where test.fseqboot contains 1000 boot strap datasets produced by 
> fseqboot).
> Is this not yet implemented in the EMBASSY version or am I overseeing something?

-M is a phylip option. All the options in PHYLIP are single letters, which do 
not work on the EMBOSS command line ... so you have to match the EMBOSS 
command line options with the PHYLIP menus.

There are some PHYLIP options, M is one of them, which we do not need. M tells 
phylip to use multiple datasets from the input. The EMBOSS version fdnadist 
will read the datasets automatically, and will set M to the number of sequence 
sets (or number of weights).

However ... I just tested this with the current version of fdnadist and found 
a bug on our Tru64 system. I suspect this was because I used the small test 
datasets that are provided in the PHYLIP documentation, but I am still 
investigating.

Please try with your dataset and let me know if you have any problems.

If anyone has some larger test datasets for the PHYLIP programs I would be 
very interested in trying fdnadist and the other programs to check we aer 
agreeing with the results from the native PHYLIP package.

EMBOSS includes a beta release of PHYLIP 3.6 - we still have to update to the 
code changes in the most recent version. We plan to do this in the enxt few 
months. More bug reports will mean we do it faster - to make sure we have the 
latest code before bug hunting :-)

Hope this helps,

Peter Rice





More information about the EMBOSS mailing list