Next: Defining explicit weights Up: Examples Previous: Aligning without additional

Secondary structure dependent gap_penalties

Example: Aligning with an alignment derived from three-dimensional structure superposition but including secondary structure dependent gap_penalties.

Secondary structure dependent penalties are included by inserting range delimiting characters into the block_file. This is illustrated by the block_file'bash_vg.bloc'. A start of range is indicated by the ``>'' character following the alignment, an end of range by the '#' character.

A typical command file that takes advantage of this information is 'bash_mult_vg.com'. In this example, the delimited ranges correspond to the conserved alpha helices in the proteins, the gap penalty is therefore INCREASED within the ranges. (An alternative strategy would be to define the ranges as the non-helix bits and then decrease the gap penalty)

Command file: bash_mult_vg.com


output_file=bash_mult_vg.out
matrix=file=ampsdir:md.mat
gap_penalty=8.0
constant=8
gap_factor=100
print_vertical=
block_file=bash_vg.bloc,1
seq_file=myoglobins.seq

The additional command is gap_factor. This defines a value by which the standard gap_penalty is multiplied WITHIN the defined ranges. Clearly, a number greater than one decreases the likelihood of a gap within the range, a number less than one increases the likelihood of a gap.

In this example there is little advantage in this approach since the alignment is very good without the use of secondary structural dependent gaps. When aligning less similar sequences, the benefits become more apparent.


gjb@bioch.ox.ac.uk