NPS@ help : sequence database help

When you see in NPS@ the sentence "Pattern value", you must paste a pattern with PROSITE syntax as indicated below.

What is the PROSITE syntax ?
You can see below a PROSITE entry.

   ID   LY6_UPAR; PATTERN.
   AC   PS00983;
   DT   JUN-1994 (CREATED); JUN-1994 (DATA UPDATE); JUL-1998 (INF UPDATE).
   DE   Ly-6 / u-PAR domain signature.
   PA   [EQR]-C-[LIVMFYAH]-x-C-x(5,8)-C-x(3,8)-[EDNQSTV]-C-{C}-x(5)-C-
   PA   x(12,24)-C.
   NR   /RELEASE=36,74019;
   NR   /TOTAL=28(20); /POSITIVE=28(20); /UNKNOWN=0(0); /FALSE_POS=0(0);
   NR   /FALSE_NEG=2; /PARTIAL=0;
   CC   /TAXO-RANGE=??E?V; /MAX-REPEAT=3;
   CC   /SITE=2,disulfide; /SITE=5,disulfide; /SITE=7,disulfide;
   CC   /SITE=10,disulfide; /SITE=13,disulfide; /SITE=15,disulfide;
   DR   P51447, CD59_AOTTR, T; P46657, CD59_CALSQ, T; Q28216, CD59_CERAE, T;
   DR   Q00996, CD59_HSVSA, T; P13987, CD59_HUMAN, T; Q28785, CD59_PAPSP, T;
   DR   P27274, CD59_RAT  , T; Q14210, E48A_HUMAN, T; Q99445, GML_HUMAN , T;
   DR   P05533, LY6A_MOUSE, T; P09568, LY6C_MOUSE, T; P35460, LY6F_MOUSE, T;
   DR   P35461, LY6G_MOUSE, T; P35459, THYB_MOUSE, T; P35457, UPAS_MOUSE, T;
   DR   P49616, UPAR_RAT  , T; P51573, UPAS_RAT  , T; Q05588, UPAR_BOVIN, T;
   DR   Q03405, UPAR_HUMAN, T; P35456, UPAR_MOUSE, T;
   DR   P47777, CD59_SAISC, N; Q64253, LY6E_MOUSE, N;
   3D   1CDQ; 1CDR; 1CDS; 1ERG; 1ERH;
   DO   PDOC00756;

The pattern is in lines starting with PA and its value is in this example :
[EQR]-C-[LIVMFYAH]-x-C-x(5,8)-C-x(3,8)-[EDNQSTV]-C-{C}-x(5)-C-x(12,24)-C.

The PROSITE syntax used :

The standard IUPAC one-letter codes.
`x' : any amino acid is accepted at the position.
`[]' : residues indicated in square parentheses are those allowed at the position.
`{}' : residues indicated in curly brackets are those forbidden at the position.
`()' : repetition of a pattern element are indicated in parenthesis.The number of repetition is indicated by a number or two number separated by a ','.
`-' : separates each pattern element.
`‹' : indicated a N-terminal restriction of the pattern.
`›' : indicated a C-terminal restriction of the pattern.
`.' : the period ends the pattern.

The x(n,m) are very time consumer so try to avoid them.
If not possible, you can divide your pattern. Start with a shorter one included in the full pattern you want to search. Then, on the subset, you can search for the full one. In our pattern example, we can search as first pattern : [EQR]-C-[LIVMFYAH]-x-C-x(5,8)-C-x(3,8)-[EDNQSTV]-C-{C}, extract full sequences containing this pattern and then search for the full pattern on this subset. Searching the full pattern on SWISS-PROT takes nearly 24 minutes. Doing the search in two steps takes 1'14 minutes (1st step takes 1'11 minute and 2nd step 3 seconds)...

So, our pattern example could be "translated" in :
[EQR]-C-[LIVMFYAH]-x-C-x(5,8)-C-x(3,8)-[EDNQSTV]-C-{C}-x(5)-C-x(12,24)-C.

   1st element [EQR]      : I want E, Q or R
   2nd element C          : I want only C
   3rd element [LIVMFYAH] : I want L, I, V, M, F, Y, A, or H
   4th element x          : I want any residue
   5th element C          : I want only C
   6th element x(5,8)     : I want any residue for 5 to 8 times
   7th element C          : I want only C
   8th element x(3,8)     : I want any residue for 3 to 8 times
   9th element [EDNQSTV]  : I want E, D, N, Q, S, T, or V
  10th element C          : I want only C
  11th element {C}        : I want all residue but C (I don't want C)
  12th element x(5)       : I want any residue 5 times
  13th element C          : I want only C
  14th element x(12,24)   : I want any residue for 12 to 24 times
  15th element C          : I want only C

Some patterns matching the syntax : 
  CD59_AOTTR:
     Q-C-Y-S-C-PYPTTQ-C-TMTT-N-C-T-SNLDS-C-LIAKAGSRVYYR-C
     Q-C-Y-S-C-PYPTTQ-C-TMTT-N-C-T-SNLDS-C-LIAKAGSRVYYRCWKFED-C
  
  GML_HUMAN :
     R-C-H-D-C-AVINDFN-C-PNIR-V-C-P-YHIRR-C-MTISIRINSRELLVYKN-C
     R-C-H-D-C-AVINDFN-C-PNIR-V-C-P-YHIRR-C-MTISIRINSRELLVYKNCTNN-C