Instruction for the command 'fkcombu' in the "KCOMBU" program

May 12, 2015
Takeshi Kawabata
(kawabata@protein.osaka-u.ac.jp)
Laboratory of Protein Informatics
Institute for Protein Research, Osaka University

[Reference for KCOMBU]

Kawabata T. Build-up algorithm for atomic correspondence between chemical structure. J.Chem.Info.Model., 2011, 51, 1775-1787.
Kawabata T., Nakamura H. Similarity-based docking using 2D maximum common substructure: dependence of prediction accuracy on target-reference chemical similarity. J.Chem.Info.Model., 2014, 54, 1850-1863.

Installation

The source code of the kcombu is mainly written in C, and developed and executed the linux environment. Some additional programs for 2D graphics and statistical analyses are written in python. For the installation, you need the gcc compiler. If you want to use another compipler, please change the "Makefile" in the "src" directory. The standard installation procedures are as follows:

download the file "kcombu-src-[date].tar.gz" file.
```
tar zxvf kcombu-src-[date].tar.gz
```
```
cd src
```
```
make -f Makefile.fkcombu
```

If the sources is successfully compiled, an execution file "fkcombu" will appear in the "../src" directory.

'fkcombu' : for flexible transforming of a target molecule onto a reference molecule

The program fkcombu flexibly superimpose the target molecule onto the reference molecule. The fkcombu program uses the terms "target" and "reference" in following definitions:

target : a compound, whose 3D conformation will be predicted by fkcombu.
reference: a compound, whose 3D conformation in binding state was known by experiment.

We assume that 3D confomrations of both reference and target molecules are available as the inputs: the 3D conformation of the reference should be experimentally determined, whereas that of target molecule is provided by a conformer generating program. The use of 3D conformation of the receptor protein is optional. 3D structure of the target molecule can be predicted by superimposing and deforming the atoms of target molecule onto that of the tamplate molecule. During the tranforming, bond lengths and bond angles of the target are kept as that of its initial conformation. Therefore, not only a good 3D conformation of the reference, but also a good 3D initial conformation of the target is necessary for the fkcombu program. The program 'fkcombu' mainly uses atomic correspondences obtained by 2D-maximum common substructure (MCS). The target molecule is transformed to superimpose each target atom onto its corresponding reference atom. The details of the calculation of MCS is described in README_pkcombu.html and README_mcs.html.

Simple Usage

Transforming of the target molecule onto the reference

  $fkcombu -T [target molecule] -R [reference molecule] -opdbT [output target in PDB]

Transforming of the target molecule onto the reference using TD-MCS1.

  $fkcombu -T [target molecule] -R [reference molecule] -opdbT [output target in PDB] -con T -mtd 1

Transforming of the target molecule onto the reference avoiding the receptor.

  $fkcombu -T [target molecule] -R [reference molecule] -P [receptor molecule] -opdbT [output target in PDB]

Assigning higher number of '-nini' (number of initial conformations) leads to a lower energy with a larger computation.
```
 $fkcombu -nini 20 -T [target molecule] -R [reference molecule] -P [receptor molecule] -opdbT [output target in PDB]
```

Transforming of the target molecule onto the reference with volume-overlap by rigid fitting

  $fkcombu -T [target molecule] -R [reference molecule] -opdbT [output target in PDB] -E V -S R

Transforming of the target molecule onto the reference with volume-overlap by flexible fitting

  $fkcombu -T [target molecule] -R [reference molecule] -opdbT [output target in PDB] -E V -S F

for showing options in detail
```
  $fkcombu -h 
```
for showing options for MCS
```
  $fkcombu -hmcs
```

Procedures for transforming

The procedures of program 'fkcombu' consists of following several steps:

[1] Finding one-to-one atomic correspondence:
The details of the calculation of MCS and its options are described in README_pkcombu.html and README_mcs.html. The default setting is Connected MCS ( -con C ). For dissimilar compound pairs, we recommended TD-MCS (-con T -mtd 1).
[2]Transform target molecule by stamping:
Executions of the stamping process can be assigned by the options -rgmch, -cfstp, -rgpca, and -cfstprng.
- -rgmch: rigid-body rmsd-min rigid-body fitting of matched atoms bwn molT and molR (T or F) [T]
- -cfstp: Conformational Change by stamping rotatable bond into molT from molR (T or F) [T]
- -cfesc: Conformational Change by minimize using only Eselfcrash (recommend for volume-flexible fitting) [F]
- -cfstprng: Stamp nonplanar 5- or 6-ring conformations (T or F) [T]
- -rgpca : rigid-body PCA-Evolmovlap fitting (T or F) [F]
[3]Transform target molecule by Gradient-based steepest descent method:
A steepest descent minimization of an potential energy is performed for initial random configurations generated by changing torsion angles randomly, which are not determined by stamping processes. A execution of the stamping process can be assigned by the option -SD,
Number of random initial configuration is assigned by the option -nini. Four potential energy functions are employed; Eatommatch, Eselfcrash, Ercptcrash and Evolmovlap.
- Eatommatch: an energy for superimposing 3D postions of MCS-corresponding atoms.
- Eselfcrash: an energy for penelizing atomic crashes between two atoms in the target molecule.
- Eprotcrash: an energy for penelizing atomic crashes between an atom in the target molecule and an atom in the receptor protein. This is optional, only works when the receptor protein (-P option) is assigned.
- Eprotatrct: an energy for attraction between an atom in the target molecule and an atom in the receptor protein. This is optional, only works when the receptor protein (-P option) is assigned. The weight -wepa should be less than the weight -wepc.
- Evolmovlap: an energy for non-specifict superimposing 3D positions of atoms using Gaussian distribution functions. This is optional.
- Etpdisrest: an energy for the Upper-distance restraint between the specified target atom and the specified protein atom. This is optional. The restraint atom pairs and upper distance is assigned by -tppair option.
Weights for these energies can be assigned by options -weam, -wesc, -wepc, -wepa, -wevo, and -wetp.

Options

Strategy options

Because options of transforming procedures and energy functions are complicated, we prepare two simple options -E and -S for controlling the details of transformation.


 -E  : Energy. 'A'tom-match, 'V'olume-overlap,         'X':by detailed options.[A]
 -S  : Search. 'F'lexible,'R'igid,      'N':do nothing 'X':by detailed options [F]

Relationships between the strategy options and other options are summarized as the following table:

Strategy Automatically assigned options

Description Options -rgdflx -mcs -cfstp -cfesc -rgmch -rgpca -SD -cfrnd -rgrnd -weam -wevo

Atom-match, Flexible -E A -S F F T T T T T 1 0

Atom-match, Rigid -E A -S R R T T T 1 0

Volume-overlap, Flexible -E V -S F F T T T T T 0 1

Volume-overlap, Rigid -E V -S R R T T T T 0 1

Input options for 'fkcombu'

Input options for 'fkcombu' are similar to those of 'pkcombu' program.

< input options for 'fkcombu'>
 -T   : target    molecule T (molT)(*.sdf|*.mol2|*.pdb|*.kcf)[]
 -R   : reference molecule R (molR)(*.sdf|*.mol2|*.pdb|*.kcf)[]
 -P   : protein   receptor P (molP)(*.sdf|*.mol2|*.pdb|*.kcf)[]
 -fT, -fR, -fP : file formats.'P'db,'S'df,'K'cf, '2':MOL2 [---]
 -aT, -aR, -aP : AtomHetero types. 'A'tom 'H'etatm 'B'oth for PDB[BBA]
 -chT,-chR,-chP: ChainIDs for PDB [---]

Options for transformation procedures

 -mcs     : Do MCS. ('T' or 'F'). If 'F', do not use any atom match [T]
 -SD      : Gradient-based Steepest Descent fitting (T or F) [T]
 -rgdflx  : Rigid or Flexible.  R'igid-body(do not set up rotational bond),'F':lexible (set up rotational bond) [F]
 -cfstp   : Conformational Change by stamping rotatable bond into molT from molR (T or F) [T]
 -cfesc   : Conformational Change by minimize using only Eselfcrash (recommend for volume-flexible fitting) [F]
 -cfstprng: Stamp nonplanar 5- or 6-ring conformations (T or F) [T]
 -rgmch   : rigid-body rmsd-min rigid-body fitting of matched atoms bwn molT and molR (T or F) [T]
 -rgpca   : rigid-body PCA-Evolmovlap fitting (T or F) [F]

Output options for transformed molecular file

 -KV      : key-value-style stdout. (T or F)[T]
 -nout    : num. of output conformations [1]
 -opdbT   : output PDB file  for molT []
 -osdfT   : output SDF file  for molT []
 -omol2T  : output MOL2 file for molT []
 -opdbTR  : output PDB file  for rotated molT and fixed molR[]
 -opdbTRP : output PDB file  for rotated molT and fixed molR and receptor protein[]
 -nameT   : new name for molecule A []
 -newresT : new residue name for molT ([resi_name]:[chain]:[resi_number]) for PDB output[: :]

Options for transformations

 -chsp3  : Chilarity fit for sp3  atoms (T or F) [F]
 -ste2D  : chirarity change for adjusting stereo parities of    molR 2D SDF file(T or F)[F]
 -hyop   : Hydrogen-position optimize (T or F)[T]

Options about multiple types of MCS atom matches

 -mcs[x] : for [x]-th type for multiple MCS matches (x=0,1,2,..). [con]:[mtd]:[at]:                                           
           ex) -mcs0 C:-1:K -mcs1 T:1:E -mcs2 T:1:K                                                                           
 -mxam   : maximum number of atom matches for the 3D-modelling [1]                                                            
 -sort : sort value for multiple target 3D models. 'E':Etotal, 'V':tanimoto_volume [E]

Detailed options for gradient optimization

 -nini  : num. of random initial conformatinon. (nini==1, only keep the input conf). [10]
 -nout  : num. of output conformations [1]
 -irb   : initial maximum step for rotational bond (degree) [180.000000]
 -iro   : initial maximum step for rotation        (degree) [180.000000]
 -ich   : initial chirarity change ('T' or 'F') [F]
 -pflp  : probabality of flip ring fragment for initial conf(0.0..1.0). [0.000000]
 -nc10  : if smallest Nselfcrash > [nc10], then [nini] x= 10. [-1]
 -xtr   : maximum step for translation     for optimization(angstrom)  [0.100000]
 -xro   : maximum step for rotation        for optimization(degree)    [10.000000]
 -xrb   : maximum step for rotational bond for optimization (degree)[10.000000]
 -nig   : Number of iteration for gradient-based fitting [100]
 -weam  : Weight for Eatommatch [1.000000]
 -wesc  : Weight for Eselfcrash [1.000000]
 -wepc  : Weight for Eprotcrash [1.000000]
 -wepa  : Weight for Eprotatrct [0.000000]
 -wevo  : Weight for Evolmovlap [0.000000]
 -wetp  : Weight for Etpdisrest [0.000000]
 -tolc  : tolerant distance for crash [1.000000]
 -dxc   : maximum distance for crash energy [6.000000]
 -fixstp: fixing rule for stamped bond for steepest descent initials. 'F':fix. 'R'otatable. [R]
 -fixpln: fixing rule for plane   bond for steepest descent initials. 'F':fix. 'I'nversion(allow 180 degree rotation). 'R'otatable.[I]
 -pvo   : parameter p for Evolmovlap [2.828427]
 -avo   : atom match type for Evolmovlap. 'I'gnore,'C'are [I]
 -tppair[x]: [x]-th target-protein atom pairs for Etpdisrest. (Tarom_num):(Patom_num):(Dupper) [x] is 0 or 1 or 2. []

options for stereo parity adjusting '-ste T'

 -pyrd : Do pyramid inversion for stereo-parity agreement (-ste) (T or F). [T]
 -arflt: Do all-atom reflection when Nchiral_disagree_impossible > 0 (T or F) [T]
 -fhxrg: fold hexa-ring with 2-disagree atoms (T or F)[T]
 -miflp: minimum angle for flip of fragment for -pflp>0.0. [10.000000]

Options for calculating MCS
Input options for calculating MCS is the same as that of 'pkcombu' program.
```
-alg : Algorithm. 'B'uild-up, 'X':eXact,'P'ca-3D-fit '3'D(raw_fit) 'F':from file (-iam option).[B]
-iam    : input file for atom matching []
```
User-defined atom matching can be used for fkcombu program. File format of atom matching is described in README_pkcombu.html.
A following command generates the transformed target molecule by the user-defined atom matching file assgined by (-iam).
```
fkcombu -T [target_molecule_file] -R [reference_molecule_file] -alg F -iam [atom_matching_file]
```
If you use the option (-alg B) as follows, the fkcombu tries to add more atom mathing to the user-defined matching using the build-up algorithm. The user-defined matching is used as the initial matching of the build-up.
```
fkcombu -T [target_molecule_file] -R [reference_molecule_file] -alg B -iam [atom_matching_file]
```
When you use the option -alg B with -iam, the user-defined matching should be satisfied the condition of MCS. Check the other options of MCS, such as -at,-con, and -mtd, described in README_pkcombu.html.