AFRC TRAINING MANUAL - CHAPTER 9
                  9.   FINDING SEQUENCE MOTIFS
                                
9.1. Locating partial sequences
     
FINDPATTERNS  locates  short  sequence  patterns,  or   'motifs'.
Pattern(s)  are specified by typing them in at the  keyboard,  or
by creating a file called PATTERN.DAT.
          
          $ Findpatterns
          
           FINDPATTERNS in what sequence(s) ?  Platelet.seq
            .
            .
                          Pattern 1:  GGAGGA
                          Pattern 2:  TTCTTC
                          Pattern 3:
          
           What should I call the output file (* Platelet.Find
          *) ?
          
           Submit Findpatterns job to which batch queue (* LONG
          *) Short

Examine the file PLATELET.FIND.

When  searching for several different patterns, it may be easier
to  give a file of patterns. FINDPATTERNS recognises the  "local
data file" name of PATTERN.DAT. An example can be fetched:
   
   $ FETCH pattern.dat
   
   $ Type pattern.dat
     
     Name    Offset  Pattern             Overhang  Documentation
     ..
     
     BamHI        1  GGATCC                     0  !
     EcoRI        1  GAATTC                     0  !
     Promotor     1  TAATA(N){20,30}ATG         0  !
     
These  files are similar to the enzyme data files used  for  the
mapping  programs and FINDPATTERNS can even read an enzyme  data
file directly.

Edit in a real or imaginary pattern of your own. This time allow
for a single mismatch, in any pattern, by using /mismatch:
   
   $ Findpatterns/mismatch=1
   
    FINDPATTERNS in what sequence(s) ?  @myoglobin.strings
   
    Search patterns read from "Pattern.Dat"
   
    What should I call the output file (* FindPatterns.Find *) ?
   myoglobin.find
   
    Submit Findpatterns job to which batch queue (* LONG *) ?
   Short

The   file  called  MYOGLOBIN.STRINGS,  was  created  using  the
STRINGSEARCH  program, identifying all entries with  "myoglobin"
in the description line. MYOGLOBIN.FIND shows all the occurences
of the patterns (with up to one mismatch) in the sequences named
in Myoglobin.strings.

9.2. Protein motifs
     
The  PROSITE  database has about 500 protein motifs,  identified
and documented from the Swissprot protein database. A motif is a
partial  sequence  associated  with  a  particular  function  or
structure  that has been identified in a group of proteins.  The
GCG  program  MOTIFS  searches for  PROSITE  patterns  within  a
sequence or sequences.
   
   $ Motifs
   
    MOTIFs from what protein sequence(s) ?  sw:kad1_human
   
    What should I call the output file (* Kad1_Human.Motifs *)
   ?
   
             Kad1_Human len:        194 .............

The  output  file  shows  the  motif,  the  fitted  pattern  and
documentation  for  each matching motif in  PROSITE.  Frequently
found patterns such as post-translational modifications are  NOT
shown unless the /FREQ command line parameter is used.


9.2.1.    Retrieving PROSITE documentation
     
MOTIFS  is  very  like FINDPATTERNS in that it uses  a  file  of
patterns (PROSITE.PATTERNS). To see what patterns are available,
retrieve the pattern file.

   
   $ FETCH prosite.patterns

Have a look at the file using the editor.

     PROSITETOGCG of: D16:[Flat.Prosite]Prosite.Doc and
     D16:[Flat.Prosite]Prosite.Da*
     
     Release 10.1  (4/93)
     
     Name            Offset Pattern                   ..
     PDoc_Name
     
     11s_Seed_Storage     1 NGx(D,E)2x2C(S,T)
     0284.PDoc
       .
       .
     Adenylate_Kinase     1 (L,I,V,M,F,Y,W)3DG(F,Y)PRx3(N,Q)
     0104.PDoc

The  11s_Seed_Storage  pattern is the first  in  the  file.  The
documentation  for that pattern is in the file  284.PDOC,  which
can also be fetched:
   
   $ FETCH 0284.Pdoc

Again, have a look at the file using the editor (or $ Type).