A set of empirical scoring functions for predicting Receptor-Ligand binding affinities (in pKd units): protein-DNA/RNA, protein-small molecule, and protein-protein complexes. Each scoring function was built using generalized linear models (GLMs) applied to specific and curated data sets.

• GCC 7.3.0 (http://gcc.gnu.org).
• R with GLM package installed (http://cran.r-project.org).

Use github to download (recommmended):

git clone git@github.com:Klab-Bioinfo-Tools/GLM-Score.git

Or if you don't have github installed:

wget https://github.com/Klab-Bioinfo-Tools/GLM-Score/archive/master.zip -O GLM-Score.zip
unzip GLM-Score.zip

cd GLM-Score-master
g++ GLM-Score_0.15.2.cpp  -o GLM-Score
cat R/small_molecule/bestGLM.Rdata-a[abc] > R/small_molecule/bestGLM.Rdata

This is an example of how to run GLM-Score:

./GLM-Score ./examples/protein.pdb ./examples/ligand.pdb ./examples/ligand.mol2 DNA

There are 4 input parameters for GLM-Score:

1. Protein or receptor structure in PDB format (www.pdb.org);
2. Ligand structure in PDB format (the ligand can also be another protein);
3. The same ligand structure, but in MOL2 format. There are many stand alone and online tools available for converting PDB files into MOL2 format. Some examples are OpenBabel (http://openbabel.org) and the online molecular formats converter (http://www.webqc.org/molecularformatsconverter.php).
4. The type of ligand molecule (DNA, protein, or small_molecule). Dependending on the type of molecule you provide, the program will choose the best statistical model for predicting the binding affinity of your binding complex.

The predicted pKD is shown on the standard output (printed on the terminal screen). For example:

./GLM-Score ./examples/protein.pdb ./examples/ligand.pdb ./examples/ligand.mol2 DNA
3.92735

The command above returned 3.92735 as the predicted pKD.

Additional results are saved in 3 files:

1. bonds.log sotres the hydrogen bond acceptor/donor pairs information:

head bonds.log

Hydrogen bonds found: 7

ATOM   1002  N   SER A  65      19.354   8.232  25.60
HETATM    4   O  BLK            16.520   8.515  25.940

ATOM    202  NH1 ARG A  13      15.904   6.417  21.320
HETATM    9   OXTBLK            16.578   8.524  23.744

ATOM   1011  OG  SER A  65      18.347   6.550  23.517
HETATM    9   OXTBLK            16.578   8.524  23.744

(continues)...

2. <ligand>_H-Bonds.pdb is a PDB file containing the identified binding site of the protein receptor plus the ligand structure.

head ligand_H-Bonds.pdb

ATOM    184  N   ARG A  13      14.993   1.315  25.032
ATOM    185  H   ARG A  13      14.308   1.698  25.683
ATOM    186  CA  ARG A  13      16.152   2.121  24.667
ATOM    187  HA  ARG A  13      16.361   1.835  23.636
ATOM    188  C   ARG A  13      17.452   1.819  25.406
ATOM    189  O   ARG A  13      18.432   2.551  25.256
ATOM    190  CB  ARG A  13      15.815   3.611  24.771
ATOM    191  HB1 ARG A  13      15.327   3.790  25.729
ATOM    192  HB2 ARG A  13      16.744   4.180  24.726
ATOM    193  CG  ARG A  13      14.896   4.098  23.662
(continues)...

3. <ligand>.interaction_terms.txt stores all predictor features generated by GLM-score: total hydrophobic contact score (V2), Van der Waals interactions (V3), side chain rotation (V4), hydrogen bonding (V5), assessible to solvent area ASA of protein (V6) and ligand (V7), repulsive interactions (V18), london disperson forces (V19), contact hydrophobicity (V20), total hydrophobicity (V21), contact surface tension (V22), total surface tension (V23)

head ligand.interaction_terms.txt

V2      V3      V4      V5      V6      V7      V18     V19     V20     V21     V22     V23    
1016.08 -797.541        8       7       17156.9 735.1   103     33.4077 3.64    91.27   335     11760.7