
PFRMAT SS  
TARGET T0110
AUTHOR 2095-5821-3749
METHOD Calculations were done using the "sausage" program
METHOD which is described in Huber, Russell, Ayers and Torda,
METHOD (1999), Bioinformatics, 15, 1064-1065.
METHOD This used a two step threading approach.  The first
METHOD score function (Huber and Torda, Protein Sci, 7, 142-149) is
METHOD for alignment to each member of a library.  This does not use
METHOD the frozen approximation.  The second score function (Huber and
METHOD Torda, Protein Sci, 7, 142-149) is built by optimising the difference
METHOD between native and non-native structure scores.
METHOD Specific parameters for this calculation are rather voluminous,
METHOD so they are listed under remarks.
METHOD 
METHOD In the above calculations, sequence residues are aligned to sites
METHOD in template structures and residue scores calculated. These scores
METHOD may be normalised for the residue type by converting to z-scores
METHOD over the set of sites to which that particular residue aligns. Gaps
METHOD are omitted from the z-score process. This allows a matrix of 
METHOD residue scores to be compiled. DSSP assignments of secondary structure
METHOD for the alignment sites was also compiled into a matrix of the same
METHOD size. It is expected that some correlation between native secondary
METHOD structure for the residue, the residue's z-score at that site, the 
METHOD secondary structure of the site and the z-score of the alignment of
METHOD the whole sequence across the library.
METHOD
METHOD Each element in this matrix is used to form a triple with one non-zero
METHOD element that is a function of the residue and alignment z-scores.
METHOD The index of the non-zero element corresponds to the site's secondary
METHOD structure. A linear model was constructed by some mathematicians that 
METHOD builds a 3x3 matrix, with which each triple is multiplied. The products
METHOD are summed to give a triple which is the secondary structure assignment
METHOD for that residue. The elements are chosen to minimise the 
METHOD misclassification rate of the matrix over a set of training matrices
METHOD derived from sequences for which the DSSP secondary structure assign-
METHOD ment is known. This method clearly generalises to a window of any 
METHOD size, and the model used here had a window size of three.
METHOD
METHOD A neural network was trained on the output triples to produce a 
METHOD single-state assignment which is used for CASP4 prediction.
MODEL 2
REMARK Tue Sep 12 18:39:25 2000
REMARK This software is guaranteed 100% free of any Biosym, MSI,
REMARK Oxford Molecular or other corporate influence.
REMARK ========================================================================
REMARK About to look for version idents
REMARK /rsc/appenzeller/u5/mabraham/../torda/bin/sausage:
REMARK      $Id: main.c,v 1.16 1999/02/15 03:53:15 arussell Exp $
REMARK      $Id: commands.c,v 1.46 1999/03/05 01:02:30 torda Exp $
REMARK      $Id: global.c,v 1.3 1998/08/25 07:06:14 torda Exp $
REMARK      $Id: misc.c,v 1.16 2000/03/14 06:07:04 arussell Exp $
REMARK      $Id: mprintf.c,v 1.22 1999/03/03 23:36:36 torda Exp $
REMARK      $Id: s_cmd.c,v 1.45 2000/08/03 08:55:12 arussell Exp $
REMARK      $Id: s_obj.c,v 1.127 2000/08/15 07:13:11 arussell Exp $
REMARK      $Id: strtools.c,v 1.5 2000/08/10 11:58:08 arussell Exp $
REMARK      $Id: vars.c,v 1.47 2000/08/15 07:14:24 arussell Exp $
REMARK      $Id: vrsn_cmd.c,v 1.1 1996/03/11 11:46:59 torda Exp $
REMARK      $Id: waste.c,v 1.1 1999/02/26 05:56:03 torda Exp $
REMARK      $Id: timer.c,v 1.4 2000/08/10 11:59:24 arussell Exp $
REMARK      $Id: sequence.c,v 1.8 2000/08/03 08:59:14 arussell Exp $
REMARK      $Id: seqgroup.c,v 1.6 2000/08/03 08:58:11 arussell Exp $
REMARK      $Id: mystring.c,v 1.6 2000/08/03 08:34:08 arussell Exp $
REMARK      $Id: seqgaps.c,v 1.6 2000/08/10 11:57:41 arussell Exp $
REMARK      $Id: aminoacids.c,v 1.5 2000/08/03 08:24:32 arussell Exp $
REMARK      $Id: readmsf.c,v 1.8 2000/08/10 12:29:11 arussell Exp $
REMARK      $Id: msa.c,v 1.7 2000/08/15 07:09:45 arussell Exp $
REMARK      $Id: alignment.c,v 1.3 2000/08/03 08:23:25 arussell Exp $
REMARK      $Id: aligngroup.c,v 1.3 2000/08/10 11:41:51 arussell Exp $
REMARK      $Id: alignmaker.c,v 1.5 2000/08/15 07:09:16 arussell Exp $
REMARK      $Id: histogram.c,v 1.5 2000/08/10 11:45:52 arussell Exp $
REMARK      $Id: simmat.c,v 1.7 2000/08/29 07:17:35 arussell Exp $
REMARK      $Id: aa_conv.c,v 1.24 1999/02/15 03:46:40 arussell Exp $
REMARK      $Id: e_malloc.c,v 1.11 1998/11/03 02:38:31 torda Exp $
REMARK      $Id: ff_sel.c,v 1.18 2000/08/10 22:52:37 torda Exp $
REMARK      $Id: readprof.c,v 1.8 2000/08/29 07:16:02 arussell Exp $
REMARK      $Id: getcoord.c,v 1.56 2000/02/27 06:50:56 torda Exp $
REMARK      $Id: getparam.c,v 1.13 2000/08/03 08:30:04 arussell Exp $
REMARK      $Id: mgc_num.c,v 1.5 1998/05/07 03:32:51 torda Exp $
REMARK      $Id: myrand.c,v 1.13 1998/01/08 00:23:42 torda Exp $
REMARK      $Id: pdbout.c,v 1.37 2000/07/20 04:35:08 torda Exp $
REMARK      $Id: read_sec.c,v 1.15 2000/08/15 00:57:27 arussell Exp $
REMARK      $Id: reg_str.c,v 1.11 1998/08/08 06:25:24 torda Exp $
REMARK      $Id: repl_str.c,v 1.7 1999/07/21 07:18:51 arussell Exp $
REMARK      $Id: s_s_alin.c,v 1.114 2000/08/29 07:17:13 arussell Exp $
REMARK      $Id: scorefct.c,v 1.29 2000/08/15 07:16:45 arussell Exp $
REMARK      $Id: sec_s.c,v 1.16 1999/07/21 08:02:02 arussell Exp $
REMARK      $Id: zfopen.c,v 1.12 2000/08/07 03:26:53 torda Exp $
REMARK      $Id: imatrix.c,v 1.8 2000/08/10 11:47:04 arussell Exp $
REMARK      $Id: fmatrix.c,v 1.7 2000/08/03 08:29:04 arussell Exp $
REMARK      $Id: needwunsch.c,v 1.17 2000/08/21 23:21:21 torda Exp $
REMARK      $Id: aahashtab.c,v 1.3 2000/08/03 08:22:12 arussell Exp $
REMARK      $Id: cc_allat_x.c,v 1.13 2000/08/03 08:26:43 arussell Exp $
REMARK      $Revision: 1.8 $
REMARK @========================================================================
M C 1.0
A C 1.0
R H 1.0
E C 1.0
F C 1.0
K C 1.0
R C 1.0
S C 1.0
D C 1.0
R C 1.0
V C 1.0
A H 1.0
Q H 1.0
E H 1.0
I H 1.0
Q H 1.0
K H 1.0
E H 1.0
I H 1.0
A C 1.0
V H 1.0
I E 1.0
L H 1.0
Q H 1.0
R H 1.0
E H 1.0
V H 1.0
K C 1.0
D C 1.0
P C 1.0
R C 1.0
I C 1.0
G C 1.0
M C 1.0
V E 1.0
T E 1.0
V E 1.0
S E 1.0
D E 1.0
V E 1.0
E E 1.0
V E 1.0
S E 1.0
S C 1.0
D C 1.0
L C 1.0
S H 1.0
Y H 1.0
A H 1.0
K H 1.0
I H 1.0
F E 1.0
V E 1.0
T E 1.0
F E 1.0
L E 1.0
F E 1.0
D E 1.0
H C 1.0
D C 1.0
E H 1.0
M H 1.0
A H 1.0
I H 1.0
E H 1.0
Q H 1.0
G H 1.0
M H 1.0
K H 1.0
G C 1.0
L C 1.0
E C 1.0
K C 1.0
A C 1.0
S C 1.0
P C 1.0
Y C 1.0
I C 1.0
R H 1.0
S H 1.0
L H 1.0
L H 1.0
G H 1.0
K H 1.0
A H 1.0
M H 1.0
R H 1.0
L H 1.0
R H 1.0
I E 1.0
V E 1.0
P E 1.0
E E 1.0
I C 1.0
R C 1.0
F E 1.0
I E 1.0
Y E 1.0
D E 1.0
Q E 1.0
S C 1.0
L C 1.0
V H 1.0
E H 1.0
G H 1.0
M C 1.0
R H 1.0
M H 1.0
S H 1.0
N H 1.0
L C 1.0
V C 1.0
T E 1.0
N E 1.0
V E 1.0
V E 1.0
R E 1.0
E E 1.0
D C 1.0
E C 1.0
K H 1.0
K C 1.0
H C 1.0
V C 1.0
E C 1.0
E C 1.0
S C 1.0
N C 1.0
REMARK =================== End of entry for prediction competition  ===========
END


