As far as file format is concerned the program proceeds as follows.  If
the file extension of the database file is .SCA (MSDOS or VAX) or SCANFILE
(CMS) it assumes that the library file is in FASTA format (ie a header
line starting with a >, and the sequence follows on subsequent lines in
upper case).  If the file extension is anything else the program uses
the GETSEQ routine described in the paper - the database format is
determined from the first line of each entry as being either EMBL,
GENBANK, CODATA, Staden (DButil) or SCANFILE.  If it is neither of these
is reads in the file in its entirity (Blank format).  Using this
approach it is no problem to have a mixed parentage file containing
entries of individual interest extracted from the various databases and
collated into a single file.

ATTENTION: If file format is NOT SCANFILE format then entries have to be
separated by lines containing a double slash:
//


In the program the following functions are called:
 
File name = xxxxxxx.SCA      getdbseq
                               getlib  /* reads in a single sequence
                                           assuming SCANFILE format */
 
File name != xxxxxx.SCA      getdata
                               getseq /* reads in a single sequence
                                         not assuming file format */
 
GETSEQ also performs the following operations:   it forces all sequence
characters to uppercase, and it translates any u(U) to T - simplifying
scanning DNA/RNA sequence databases.  The latter is necessary for
convience until the databases are circulated all in one format (I know
EMBL has declared its intension to do so).  However since this is
preformaed using a table lookup that would have to be performed anyway
there is no additional penalty in terms of program execution.
 
As described in the paper when the motif (or pattern) is defined case is
significant.  Normal sequence characters should be given in upper case
as the lower case characters are reserved for the IUB nucleic acid
symbols: the ambiguous bases are translated by the program into the
equivalent base set (eg r becomes [AG] and y becomes [CT] ).