
This project consists of generation of a homology model for the Fv
fragment of the antibody herceptin.  The calculation demonstrates the
use of multiple templates.


POPULATE USER PROJECT DIRECTORY WITH INPUT FILES

The following files entered into directories "/up1/arg" and "/up1/seq"
were prepared by hand editing.

/up1/arg/domains
  ranges of residues within corresponding PDB entries for antibody Fv
  fragments taken from the set of SCOP95 domain definitions

/up1/arg/exptstructs.fv_kap
  templates for 201 type kappa Fv fragments

/up1/arg/exptstructs.fv_lam
  templates for 24 type lambda Fv fragments

/up1/arg/exptstructs.fv_tcr
  templates for 12 T-cell receptor Fv fragments

/up1/seq/seq.herceptin
  a sequnce format file for the target herceptin Fv fragment

For each template of file "/up1/arg/exptstructs.fv_kap", a PDB format
file is created in directory "/up1/exp" by copying the corresponding
PDB entry and editing to remove all atoms not part of the SCOP95
domain definition.

/up1/exp/12e8lh.pdb
/up1/exp/1a14lh.pdb
  .
  .
  .
/up1/exp/Aqfwmi.pdb


PROCESS TEMPLATES

For each template TEM, the "greg" command regularizes geometry.

|SYNTAX
|  greg up1 TEM
|INPUT FILES                   OUTPUT FILES
|/up1/exp/TEM.pdb              /up1/dgn/greg.TEM
|                              /up1/seq/seq.TEM
|                              /up1/tor/tor.TEM.00
|                              /up1/car/TEM.00.pdb

The RMSD between the final regularized structure and the starting
experimental structure ranges from 0.081 to 0.265 angstroms.

For each template, the "prof" command creates a profile of defect
energy density.

|SYNTAX
|  prof up1 TEM 00
|INPUT FILES                   OUTPUT FILES
|/up1/seq/seq.TEM              /up1/dgn/prof.TEM.00
|/up1/tor/tor.TEM.00           /up1/car/dprof.TEM.00
|/up1/car/TEM.00.pdb


BUILD HOMOLOGY MODEL

The "hlog" command builds a homology model for target sequence
herceptin using the group of template structures fv_kap.

|SYNTAX
|  hlog up1 herceptin fv_kap
|INPUT FILES                   OUTPUT FILES
|/up1/arg/exptstructs.fv_kap   /up1/dgn/hlog.herceptin.fv_kap
|/up1/seq/seq.herceptin        /up1/car/TEM.fv_kap.pdb
|/up1/seq/seq.TEM              /up1/tor/tor.herceptin.fv_kap
|/up1/tor/tor.TEM.00           /up1/car/herceptin.fv_kap.pdb
|/up1/car/TEM.00.pdb           /up1/car/dprof.herceptin.fv_kap
|/up1/car/dprof.TEM.00         /up1/stp/stp.hot.herceptin

Following identification of the largest template, the algorithm aligns
all other template structures to the structure of the largest
template.  In this case, the largest template is 1igmlh with 244
residues.  For each template other than the largest template,
structure-structure alignment to the largest template is restricted to
the largest subset of residues for which low RMSD alignment is
possible.  Over all structure-structure alignments, the size of the
subset of aligned residues ranges from 205 to 226 residues, the RMSD
ranges from 0.74 to 2.51 angstroms with 1 outlier, 5.66 angstroms for
the alignment of 1il1ba to 1igmlh.  Structure is conserved over all
201 templates for 13 segments containing 139 residues.

For each template, the sequence of herceptin is aligned to the
sequence of the template.  For each segment of the multiple structure
alignment, including both structurally conserved and non-structurally
conserved segments, the template for which the alignment score to the
target herceptin is highest within the segment is selected for use in
building a model.  The heavy chain of herceptin aligns best to
template 1mhplh over all structurally conserved segments.  Over
non-structurally conserved segments, 5 templates 1mhplh, 1j05ab,
1a3rlh, 1deeab, and 1cz8lh provide optimal alignments.  The light
chain of herceptin aligns best to template 1l7ilh over all
structurally conserved segments.  Over non-structurally conserved
segments, 4 templates 1l7ilh, 1ncblh, 1iailh, and 1bvkab provide
optimal alignments.

The algorithm begins the building of a model for the target sequence
by copying coordinates for aligned atoms from the multiple structure
alignment using the template selected as best for the segment.


REFINE HOMOLOGY MODEL

Finally, the "rcyc" command is used to improve the model structure by
minimization of full energy.

|FUNCTIONALITY
|  energy refinement of a homology model
|SYNTAX
|rcyc up1 herceptin fv_kap
|INPUT FILES                   OUTPUT FILES
|/up1/seq/seq.herceptin        /up1/dgn/rcyc.herceptin
|/up1/tor/tor.herceptin.fv_kap /up1/stp/stp.g???.herceptin
|                              /up1/car/herceptin.g???.pdb
|                              /up1/tor/tor.herceptin.g???

The final energy refined structure from the sequence generated by the
"rcyc" command is given a more meaningful conformation name by copying
and renaming.

cp ../../up1/tor/tor.herceptin.g056 ../../up1/tor/tor.herceptin.model
cp ../../up1/car/herceptin.g056.pdb ../../up1/car/herceptin.model.pdb

In this case, the "rcyc" command requires a large computation time.
