Next: Save_lookup format Up: MULTALIGN file formats Previous: tree_file

block_file

This defines a multiple alignment in vertical format. The print_vertical command produces a file in block_file format.


The minimum requirements for a block_file for N aligned sequences are
1.  N  '>comment line(s)'
2.  '* iteration int'
3.  'N or more vertically aligned sequences'
4.  '*'

  1. The comment lines define the sequence identifiers and the number of '>' characters preceding the first '* iteration int' line define the number of sequences that are defined in the sequence lines.

  2. This line specifies the beginning of the alignment to be read. The '*' character specifies the column in which the alignment begins. The 'iteration int' specifier identifies the particular alignment within this block_file. Several alignments may follow each other providing they are identified by a different iteration number (eg. 1,2,3).

  3. The alignment is ended by a '*' character which should be in the same column as the '*' character that started the alignment.

Simple example:


This is a block file containing two alternative alignments of three sequences.
The comments that I an writing here may appear in the block file, but are
ignored by MULTALIGN when the file is read.  The only proviso is that no
'greater than' or 'star' characters must be present.
>first  this is sequence A
>second this is sequence B
>third  This is sequence C
* iteration 1
a  
a p
avg
llg
lcr
g
 pg
www
s	
*
*iteration 2
a
a p
avg
llg
lcr
gpg
www
s
*

Examples: See bash.bloc, bash_vg.bloc and globin.bloc.

The block_filemay also encode regions to apply variable gap_penaltieswithin. See bash_vg.bloc for example syntax.


gjb@bioch.ox.ac.uk