This document describes the different types of SeqAlign's produced 
by BLAST.  

Familiarity with the toolkit documentation and the material on SeqAlign's
covered at ftp://ncbi.nlm.nih.gov/toolbox/ncbi_tools/sdkdoc/SEQALIGN.HTML
is assumed. 


ungapped blastn/blastp:
-----------------------

A dense-diag is used, one dense-diag/alignment.  All alignments for one 
database sequence are grouped into one SeqAlign.


gapped blastn/blastp:
---------------------

A dense-seg is used, one dense-seg/alignment.  A SeqAlign can only accomodate
one dense-seg, so a separate SeqAlign is used for every alignment.


ungapped blastx/tblastn/tblastx:
--------------------------------

A std-seg is used for this, as the nucleotide and protein sequences
have different lengths.  The coordinates for the translated nucleotide
sequence are those of the original nucleotide sequence.
All alignments for one database sequence are grouped into one SeqAlign.

gapped blastx/tblastn:
----------------------

A std-seg is used for this, as the nucleotide and protein sequences
have different lengths.  The coordinates for the translated nucleotide
sequence are those of the original nucleotide sequence.
There is one alignment per SeqAlign, but multiple std-seg's may
be in one alignment if there are gaps.  A gap is represented by an
empty std-seg.  The alignment below can be represented by 
three std-segs; the first describes the alignment until the gap in 
the database sequence; a std-seg with an empty second Seq-loc describes 
the gap; finally a third std-seg describes the last portion of the alignment.


Query: 787 HRSNPPTHNNSHSVYGLRPTMRTDIILGSNSYYQPPLSNP 906
           HR N  TH N H++ G+ PT    ++ G    Y+PP   P
Sbjct: 481 HRGNNGTHINGHALNGVAPT----LVNGRGKSYEPPAPPP 516