[EMBOSS] Batch alignments?

Daniel Barker db60 at st-andrews.ac.uk
Sat Jan 29 12:08:34 UTC 2011

Dear Robin,

In a Perl script, you could read your sequences using BioPerl Seq::IO, 
create input files, and execute the EMBOSS commands from Perl with 
system() or `...`. Remember to close temporary files you've created in 
Perl before using them - this forces the last bit of the file to be 
really written, rather than stay languishing in some buffer.

This is just one suggestion - there will be other ways. An excellent 
guide to Perl is at: http://korflab.ucdavis.edu/Unix_and_Perl

If I remember, some earlier version of needle had a bug which could 
affect scores. So you might want to upgrade.

For coding sequences, consider doing the alignment at the protein level 
- unless there's a good reason not to. I.e. translate your coding 
sequences (perhaps with EMBOSS transeq) and align the translations. 
Generally, this will be more sensitive and less easily misled. If you 
require nucleotide alignment, you can convert back to the original 
nucleotides afterwards - but this is more likely to be required with 
multiple alignment than with pairwise (and for multiple alignments can 
be achieved with EMBOSS tranalign).

Best regards,


Daniel Barker
The University of St Andrews is a charity registered in Scotland : No 

More information about the EMBOSS mailing list