mpialign - parallel local alignment of biological sequences
mpialign [-s scores] [-S split] [-H hblk] [-V vblk] file1
file2
mpiAlign is a local sequence aligner, especially intended for use
with large biological DNA sequences, with more than 1Mbp (Millions of base
pairs). It uses the Smith-Waterman exact algorithm with affine gap cost
function to perform this task.
- -h
- display this help and exit
- -s <scores>
- specify a comma-separated list of scores to be used while calculating
aligment matrices throughout the program. The list must be in the format
"-sMATCH,MISMATCH,GAP_OPEN,GAP_EXTENSION" (without quotes), and
is parsed in this PRECISE order; no spaces are allowed between values. If
there are any unspecified parameters, these are set to default values and
a warning message is issued; exceeding parameters are discarded
- -S <split>
- (mpialign only) number of parts in which to split the alignment matrix;
after this step a cyclic block model is applied, subdividing each part
equally between all available nodes
- -H <hblk>
- (mpialign only) number of horizontal subdivisions made by each node to its
alignment submatrix; since this value defines block width, a good choice
should allow two full matrix lines to fit the processor's cache pages,
improving algorithm performance
- -V <vblk>
- (mpialign only) number of vertical subdivisions made by each node to its
alignment submatrix; this value directly affects the amount of internode
communication and is used ONLY if 'split' is set to 1, otherwise it is set
to the number of available nodes