Program : SAVANT
Version : 060804
Author : Gerard J. Kleywegt, Dept. of Cell and Molecular Biology,
Uppsala University, Biomedical Centre, Box 596,
SE-751 24 Uppsala, SWEDEN
E-mail : gerard@xray.bmc.uu.se
Purpose : post-processing of SPASM hits
Package : SPASM
Reference(s) for this program:
* 1 * G.J. Kleywegt (1998). Deja-vu all over again. CCP4 Newsletter on Protein Crystallography 35, July 1998, pp. 10-12. [http://xray.bmc.uu.se/usf/factory_9.html]
* 2 * G.J. Kleywegt & T.A. Jones (1998). Databases in protein crystallography. Acta Cryst D54, 1119-1131. [http://xray.bmc.uu.se/gerard/papers/databases.html] [http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Retrieve&list_uids=10089488&dopt=Citation] [http://scripts.iucr.org/cgi-bin/paper?ba0001]
* 3 * G.J. Kleywegt (1999). Recognition of spatial motifs in protein structures. J Mol Biol 285, 1887-1897. [http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Retrieve&list_uids=9917419&dopt=Citation]
* 4 * D. Madsen & G.J. Kleywegt (2002). Interactive motif and fold recognition in protein structures. J. Appl. Cryst. 35, 137-139. [http://scripts.iucr.org/cgi-bin/paper?wt0007]
* 5 * Kleywegt, G.J., Zou, J.Y., Kjeldgaard, M. & Jones, T.A. (2001). Around O. In: "International Tables for Crystallography, Vol. F. Crystallography of Biological Macromolecules" (Rossmann, M.G. & Arnold, E., Editors). Chapter 17.1, pp. 353-356, 366-367. Dordrecht: Kluwer Academic Publishers, The Netherlands.
990902 - 0.1 - first version
990903 - 0.2 - ask the user for MODE; produce an O2D scatter plot
file of the hits (X=nr of matched atoms; Y=RMSD)
990907 - 0.3 - add commented-out sketch_cpk and sketch_stick instructions
for every hit in the output O macro file
990928 - 1.0 - implemented alternative residues/atoms for fuzzy
pattern matching (NOTE: also made a new version
of the "savant.lib" library file which is necessary
for this version of the program !)
991013 -1.0.1- minor change
991119 - 1.1 - SAVANT now does the least-squares superpositioning of the
hits itself before it writes the small PDB files
001112 - 1.2 - minor changes
001213 - 1.3 - major bug fix; several minor changes
040302 -1.3.1- made O macros produced by SAVANT compatible with
current O (i.e., they read stereo_chem.odb and they
use the pdb_read command instead of sam_at_in)
041001 - 1.4 - replaced Kabsch' routine U3BEST by quaternion-based
routine (U3QION) to do least-squares superpositioning
050919 -1.4.1- print number of every hit
060804 - 1.5 - various changes under the hood (synchronised with SAVANA)
SPASM is a cute program, but sometimes the sacrifice of sensitivity for the sake of speed and reduced database size may be too much. SAVANT is a post-processor for SPASM hits that will do an all-atom superpositioning of the patterns found in the PDB (provided the hits are available locally on disk).
SAVANT requires a little input file which can be created with SPASM, as well as a library (optional) of possible alternative identities of certain atom types (e.g., OD1 and OD2 in ASP). The output is a new O macro and an ODB file with the least-squares RT operators. These in turn can be run through DEJANA to sort them, and to throw away any poor matches.
SAVANT will also write all the hits to files in a sub-directory called "savant" (this subdirectory must exist, or SAVANT will refuse to run). A major benefit of this is that, since only the residues that constitute the hits are written, you can inspect many more hits in O than with the regular SPASM macros (since the latter read in the entire PDB files, which tends to fill up your database rather quickly).
In summary, reasons why you should use SAVANT after SPASM:
- better RT operators to bring the hits on top of your pattern
- easier to separate good hits from poor ones (with DEJANA)
- small PDB files are created for each of the hits, making
the O macro to compare the hits much faster (and it won't
fill up the O database as quickly)
- you can run SPASM looking for main chain plus side chain (e.g.,
if you only have two residues in your pattern), but in
SAVANT you can do the all-atom superpositioning using only
side-chain atoms if you like (as long as there are at least
three atoms to do the superpositioning)
- if you want, you can do the superpositioning on a small subset
of atoms (e.g., if you have several interacting carboxylates,
remove all side-chain atoms from the pattern PDB file, except
the oxygens, and then only these oxygens will be used by
SAVANT - but make sure to keep at least one atom for every
residue that you want to have displayed (e.g., the CA) !)
The following input is required (note that the library file is optional; if it doesn't exist the program will still run):
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- SAVANT library file ? (/home/gerard/lib/savant.lib) ../savant.lib Nr of AMB lines : ( 11) Nr of ALT lines : ( 13)SAVANT input file ? (savant.inp) cbh2.savant SAVANT O macro ? (savant.omac) SAVANT ODB file ? (savant.odb) ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
The library file may look like this:
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----! savant.lib - g j kleywegt @ 990902,28 ! ! Ambiguous atom types: ! AMB RES ATM1 ATM2 -> list ambiguous atom types ! (A3,1X,A3,1X,A4,1X,A4) ! ! Matching atom types in different residue types: ! ALT RES ATM1 RE2 ATM2 -> list possible substitutions ! (A3,1X,A3,1X,A4,1X,A3,1X,A4) ! (CB is *always* considered) ! ! NOTE: ALT stuff not implemented in SAVANT yet !!! ! !MB RES ATM1 ATM2 AMB PHE CD1 CD2 AMB PHE CE1 CE2 AMB TYR CD1 CD2 AMB TYR CE1 CE2 AMB HIS ND1 CD2 AMB HIS CE1 NE2 AMB ASP OD1 OD2 AMB ASN OD1 ND2 AMB GLU OE1 OE2 AMB GLN OE1 NE2 AMB ARG NH1 NH2 ! ! the following ones *shouldn't* be ambiguous !!! ! !AMB THR OG1 CG2 !AMB VAL CG1 CG2 !AMB ILE CG1 CG2 !AMB LEU CD1 CD2 ! !LT RES ATOM RES ATOM ALT PHE CG TYR CG ALT PHE CD1 TYR CD1 ALT PHE CD1 TYR CD2 ALT PHE CE1 TYR CE1 ALT PHE CE1 TYR CE2 ALT PHE CD2 TYR CD2 ALT PHE CD2 TYR CD1 ALT PHE CE2 TYR CE1 ALT PHE CE2 TYR CE2 ALT PHE CZ TYR CZ ! [...] ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
First the pattern you searched for is read:
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- MODE : ( 1) MC + SCPattern PDB file : (cbh2.pdb) Residues in pattern : ( 3) Atoms in residues : ( 12 8 8) Total : ( 28) Ambiguous : ( 8) ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
Then all the hits are scrutinised:
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- Hit PDB file : (/nfs/pdb/full/1b2n.pdb) Residues in hit : ( 3) Hit will go to PDB file : (savant/cbh2_1b2n_1.pdb) Atoms in residues : ( 12 8 8) Atoms in common : ( 28) RMS distance : ( 2.458) Nr ambiguous : ( 8) TYR 169 CD1 <-> CD2 => RMSD (A) : ( 2.380) TYR 169 CE1 <-> CE2 => RMSD (A) : ( 2.297) ASP 175 OD1 <-> OD2 => RMSD (A) : ( 2.330) ASP 221 OD1 <-> OD2 => RMSD (A) : ( 2.289) Lowest RMSD : ( 2.289)[...]
Hit PDB file : (/nfs/pdb/full/1cb2.pdb) Residues in hit : ( 3) Hit will go to PDB file : (savant/cbh2_1cb2_1.pdb) Atoms in residues : ( 11 8 8) Atoms in common : ( 20) RMS distance : ( 0.693) Nr ambiguous : ( 4) ASP 175 OD1 <-> OD2 => RMSD (A) : ( 0.062) ASP 221 OD1 <-> OD2 => RMSD (A) : ( 0.691) Lowest RMSD : ( 0.062) ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
Note that, if you have read the library file, and if you included side-chains in your search, any "ambiguous" atom types are tested to see if switching the atom names results in a lower RMSD. If this is the case, the switch is kept, and the next ambiguous atom type is tested.
Running the resulting hits through DEJANA may yield something like this:
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- Min nr of matched atoms/residues/SSEs ? ( 1) Max RMSD of matched atoms/residues/SSEs ? ( 999.990)Sorting hits ...
Nr of hits left : ( 69)
# 1 ID 1iph1 Nres 28 RMSD 1.73 A # 2 ID 2fok1 Nres 28 RMSD 1.76 A # 3 ID 1noy1 Nres 28 RMSD 1.79 A # 4 ID 1waj1 Nres 28 RMSD 1.83 A # 5 ID 1ctn1 Nres 28 RMSD 1.84 A # 6 ID 1sat1 Nres 28 RMSD 1.86 A
[...]
# 36 ID 1bxo1 Nres 28 RMSD 2.39 A # 37 ID 1ksi1 Nres 28 RMSD 2.42 A # 38 ID 1cb21 Nres 20 RMSD 0.06 A # 39 ID 1yge1 Nres 20 RMSD 1.23 A # 40 ID 1ctn2 Nres 20 RMSD 1.61 A # 41 ID 2hvm2 Nres 20 RMSD 1.77 A
[...]
# 68 ID 2cae1 Nres 20 RMSD 2.44 A # 69 ID 1kcw1 Nres 20 RMSD 2.56 A
Select one of the following options: 0 = re-enter criteria and re-sort 1 = write new O macro with current hits 2 = quit program without writing new O macro Option (0, 1, 2) ? ( 0) ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
Thew output may look as follows:
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- SAVANT library file ? (/home/gerard/lib/savant.lib) ../savant.lib Nr of AMB lines : ( 11) Nr of ALT lines : ( 13)SAVANT input file ? (savant.inp) cra5.savant SAVANT O macro ? (savant.omac) SAVANT ODB file ? (savant.odb)
MODE : ( 2) MC only
Pattern PDB file : (cra5.pdb) Residues in pattern : ( 9) Atoms in residues : ( 9 8 8 9 4 9 4 7 9) Total : ( 67) Ambiguous : ( 2)
Hit PDB file : (/nfs/pdb/full/1ar1.pdb) Residues in hit : ( 9) Hit will go to PDB file : (savant/cra5_1ar1_1.pdb) Atoms in residues : ( 7 5 7 8 7 8 10 5 14) Atoms in common : ( 36) RMS distance : ( 1.357)
Hit PDB file : (/nfs/pdb/full/1bgw.pdb) Residues in hit : ( 9) Hit will go to PDB file : (savant/cra5_1bgw_1.pdb) Atoms in residues : ( 14 11 11 4 14 7 4 7 8) Atoms in common : ( 36) RMS distance : ( 1.313)
Hit PDB file : (/nfs/pdb/full/1cbi.pdb) Residues in hit : ( 9) Hit will go to PDB file : (savant/cra5_1cbi_1.pdb) Atoms in residues : ( 7 8 8 9 4 8 4 7 9) Atoms in common : ( 36) RMS distance : ( 0.605)
[...]
Hit PDB file : (/nfs/pdb/full/2fhe.pdb) Residues in hit : ( 9) Hit will go to PDB file : (savant/cra5_2fhe_1.pdb) Atoms in residues : ( 8 11 4 7 6 7 6 10 7) Atoms in common : ( 36) RMS distance : ( 1.710) ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
And in DEJANA:
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- Min nr of matched atoms/residues/SSEs ? ( 1) Max RMSD of matched atoms/residues/SSEs ? ( 999.990)Sorting hits ...
Nr of hits left : ( 19)
# 1 ID 1cbs1 Nres 36 RMSD 0.00 A # 2 ID 1cbi1 Nres 36 RMSD 0.60 A # 3 ID 1yfo1 Nres 36 RMSD 0.99 A # 4 ID 1fdr1 Nres 36 RMSD 1.08 A # 5 ID 1bgw1 Nres 36 RMSD 1.31 A
[...] ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
If you have used SPASM for fuzzy pattern matching (e.g., your motif contains an Arg, but you will tolerate a Lys as well), SAVANT can (from version 1.0) take this into account, using the set of alternative residues/atoms defined in the library file (CB atoms are always considered, so they don't have to be included explicitly in the library file).
----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- SAVANT library file ? (/home/gerard/lib/savant.lib) Nr of AMB lines : ( 11) Nr of ALT lines : ( 37)SAVANT input file ? (savant.inp) cra2.savant SAVANT O macro ? (savant.omac) SAVANT ODB file ? (savant.odb) O2D plot file ? (savant.plt)
Select mode: 0 = SC, 1 = MC + SC, 2 = MC Mode (0/1/2) ? ( 1) 0 MODE : ( 0) SC only
Pattern PDB file : (cra2.pdb) Residues in pattern : ( 3) Atoms in residues : ( 11 11 12) Total : ( 34) Ambiguous : ( 8) Alternative : ( 10)
Hit PDB file : (/nfs/pdb/full/169l.pdb) Residues in hit : ( 3) Hit will go to PDB file : (savant/cra2_169l_1.pdb) Atoms in residues : ( 9 11 11) ==> Match ARG CB <-> LYS CB ==> Alternative ? ARG CG ==> Alternative ? ARG CD ==> Not matchable ARG NE ==> Not matchable ARG CZ ==> Not matchable ARG NH1 ==> Not matchable ARG NH2 ==> Match ARG CB <-> ARG CB ==> Match ARG CG <-> ARG CG ==> Match ARG CD <-> ARG CD ==> Match ARG NE <-> ARG NE ==> Match ARG CZ <-> ARG CZ ==> Match ARG NH1 <-> ARG NH1 ==> Match ARG NH2 <-> ARG NH2 ==> Match TYR CB <-> PHE CB ==> Alternative ? TYR CG ==> Alternative ? TYR CD1 ==> Alternative ? TYR CE1 ==> Alternative ? TYR CD2 ==> Alternative ? TYR CE2 ==> Alternative ? TYR CZ ==> Not matchable TYR OH Atoms in common : ( 9) RMS distance : ( 1.145) Nr ambiguous : ( 2) ARG 132 NH1 <-> NH2 => RMSD (A) : ( 1.404) Lowest RMSD : ( 1.145) Nr alternatives : ( 8) ARG 111 CG <-> LYS A 60 CG => RMSD (A) : ( 1.228) ARG 111 CD <-> LYS A 60 CD => RMSD (A) : ( 1.425) TYR 134 CG <-> PHE A 4 CG => RMSD (A) : ( 1.440) TYR 134 CD1 <-> PHE A 4 CD1 => RMSD (A) : ( 1.526) TYR 134 CD1 <-> PHE A 4 CD2 => RMSD (A) : ( 1.418) TYR 134 CE1 <-> PHE A 4 CE1 => RMSD (A) : ( 1.648) TYR 134 CE1 <-> PHE A 4 CE2 => RMSD (A) : ( 1.452) TYR 134 CD2 <-> PHE A 4 CD1 => RMSD (A) : ( 1.698) TYR 134 CD2 <-> PHE A 4 CD2 => RMSD (A) : ( 1.493) TYR 134 CE2 <-> PHE A 4 CE1 => RMSD (A) : ( 1.884) TYR 134 CE2 <-> PHE A 4 CE2 => RMSD (A) : ( 1.637) TYR 134 CZ <-> PHE A 4 CZ => RMSD (A) : ( 1.791) Final RMSD : ( 1.791) Nr matches : ( 17)
[...]
Hit PDB file : (/nfs/pdb/full/1waj.pdb) Residues in hit : ( 3) Hit will go to PDB file : (savant/cra2_1waj_1.pdb) Atoms in residues : ( 9 11 11) ==> Match ARG CB <-> LYS CB ==> Alternative ? ARG CG ==> Alternative ? ARG CD ==> Not matchable ARG NE ==> Not matchable ARG CZ ==> Not matchable ARG NH1 ==> Not matchable ARG NH2 ==> Match ARG CB <-> ARG CB ==> Match ARG CG <-> ARG CG ==> Match ARG CD <-> ARG CD ==> Match ARG NE <-> ARG NE ==> Match ARG CZ <-> ARG CZ ==> Match ARG NH1 <-> ARG NH1 ==> Match ARG NH2 <-> ARG NH2 ==> Match TYR CB <-> PHE CB ==> Alternative ? TYR CG ==> Alternative ? TYR CD1 ==> Alternative ? TYR CE1 ==> Alternative ? TYR CD2 ==> Alternative ? TYR CE2 ==> Alternative ? TYR CZ ==> Not matchable TYR OH Atoms in common : ( 9) RMS distance : ( 0.703) Nr ambiguous : ( 2) ARG 132 NH1 <-> NH2 => RMSD (A) : ( 1.239) Lowest RMSD : ( 0.703) Nr alternatives : ( 8) ARG 111 CG <-> LYS 240 CG => RMSD (A) : ( 0.738) ARG 111 CD <-> LYS 240 CD => RMSD (A) : ( 0.774) TYR 134 CG <-> PHE 266 CG => RMSD (A) : ( 0.761) TYR 134 CD1 <-> PHE 266 CD1 => RMSD (A) : ( 0.907) TYR 134 CD1 <-> PHE 266 CD2 => RMSD (A) : ( 0.984) TYR 134 CE1 <-> PHE 266 CE1 => RMSD (A) : ( 1.120) TYR 134 CE1 <-> PHE 266 CE2 => RMSD (A) : ( 1.163) TYR 134 CD2 <-> PHE 266 CD1 => RMSD (A) : ( 1.173) TYR 134 CD2 <-> PHE 266 CD2 => RMSD (A) : ( 1.094) TYR 134 CE2 <-> PHE 266 CE1 => RMSD (A) : ( 1.190) TYR 134 CE2 <-> PHE 266 CE2 => RMSD (A) : ( 1.066) TYR 134 CZ <-> PHE 266 CZ => RMSD (A) : ( 1.128) Final RMSD : ( 1.128) Nr matches : ( 17)
[...]
Hit PDB file : (/nfs/pdb/full/7aat.pdb) Residues in hit : ( 3) Hit will go to PDB file : (savant/cra2_7aat_1.pdb) Atoms in residues : ( 11 9 12) ==> Match ARG CB <-> ARG CB ==> Match ARG CG <-> ARG CG ==> Match ARG CD <-> ARG CD ==> Match ARG NE <-> ARG NE ==> Match ARG CZ <-> ARG CZ ==> Match ARG NH1 <-> ARG NH1 ==> Match ARG NH2 <-> ARG NH2 ==> Match ARG CB <-> LYS CB ==> Alternative ? ARG CG ==> Alternative ? ARG CD ==> Not matchable ARG NE ==> Not matchable ARG CZ ==> Not matchable ARG NH1 ==> Not matchable ARG NH2 ==> Match TYR CB <-> TYR CB ==> Match TYR CG <-> TYR CG ==> Match TYR CD1 <-> TYR CD1 ==> Match TYR CE1 <-> TYR CE1 ==> Match TYR CD2 <-> TYR CD2 ==> Match TYR CE2 <-> TYR CE2 ==> Match TYR CZ <-> TYR CZ ==> Match TYR OH <-> TYR OH Atoms in common : ( 16) RMS distance : ( 2.209) Nr ambiguous : ( 6) ARG 111 NH1 <-> NH2 => RMSD (A) : ( 2.212) TYR 134 CD1 <-> CD2 => RMSD (A) : ( 2.123) TYR 134 CE1 <-> CE2 => RMSD (A) : ( 2.020) Lowest RMSD : ( 2.020) Nr alternatives : ( 2) ARG 132 CG <-> LYS A 258 CG => RMSD (A) : ( 2.037) ARG 132 CD <-> LYS A 258 CD => RMSD (A) : ( 2.036) Final RMSD : ( 2.036) Nr matches : ( 18) ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE ----- EXAMPLE -----
Note that the matching of Tyr and Phe is ambiguous for several atom types (e.g., Phe CD1 can be matched with Tyr CD1 and CD2). SAVANT will use the match that gives the lowest overall RMSD. However ... SAVANT does not try all possible permutations ! This means that the same atom can sometimes be used twice (see the first example in the output above for an example: Tyr CD1 and CD2 are both matched to Phe CD2). Also, sometimes you may get "silly" matches such as CD1-CD1/CD2-CD2 with CE1-CE2/CE2-CE1. However, usually things work fine, and if this is not the case it tends to be for the poorer hits (compare the second example in the output above with the first).
None (other than the features mentioned under "fuzzy pattern matching").
Created at Fri Aug 4 21:34:56 2006
by MAN2HTML version 060130/2.0.7 . This manual
describes SAVANT, a program of the Uppsala
Software Factory (USF), written and maintained by Gerard Kleywegt. © 1992-2006.