The Hotspots API is the Python package for the Fragment Hotspot Maps project, a knowledge-based method for determining small molecule binding "hotspots".
For more information on this method:
Although the Hotspots API is publicly available, it is dependant on the CSD Python API - a commercial package.
If you are an academic user, it's likely your institution will have a license. If you are unsure if you have a license or would like to enquire about purchasing one, please contact [email protected].
Please note, this is an academic project and we would therefore welcome feedback, contributions and collaborations. If you have any queries regarding this package please contact us ([email protected])!
The CSDS is available from here.
You will need your customer number and activation key. You must activate your license before proceeding.
Ghecom is available from here.
"The source code of the GHECOM is written in C, and developed and executed on the linux environment (actually on the Fedora Core). For the installation, you need the gcc compiler. If you do not want to use it, please change the "Makefile" in the "src" directory."
Download the file ghecom-src-[date].tar.gz file.
tar zxvf ghecom-src-[date].tar.gz
cd src
make
NB: The executable will be located at the parent directory.
Download the environment.yml file from the github repositiory.
Open the file, and edit the file path to the your local ccdc conda channel that was installed as part of your CSDS:
"file:///home/pcurran/CCDC/Python_API_2020/ccdc_conda_channel"
Save and close environment.yml.
Create conda environment using the environment.yml:
conda create -n hotspots -f environment.yml
Finally, there are a few environment variables to set:
$ export CSDHOME=/home/my_ccdc_software_dir/CCDC/CSD_2020
$ export LD_LIBRARY_PATH=$CONDA_PREFIX/lib:$CONDA_PREFIX/lib/python3.7/site-packages/ccdc/_lib:$LD_LIBRARY_PATH
$ export GHECOM_EXE=$PREFIX/ghecom_latest/ghecom
We recommend saving these within your conda environment. To do this, see setup_environment.sh shell script within
the hotspots repositiory. For more details on saving environment variables, see the
conda documentation.
Install Hotspots v1.0.3:
a) Latest stable release (recommended for most users):
conda activate hotspots
pip install hotspots
or
pip install https://github.com/prcurran/hotspots/archive/v1.0.3.zip
b) Very latest code
mkdir ./hotspots_code
cd hotspots_code
git clone [email protected]:prcurran/hotspots.git
conda activate hotspots
pip install ./hotspots
Start activating your Anaconda environment and setting some variables.
conda activate hotspots
export GHECOM_EXE=<path_to_GHECOM_executable>
export CSDHOME=<path_to_CSDS_installation>/CSD_2019
The first step is to make sure your protein is correctly prepared for the calculation. The structures should be protonated with small molecules and waters removed. Any waters or small molecules left in the structure will be included in the calculation.
One way to do this is to use the CSD Python API:
from ccdc.protein import Protein
prot = Protein.from_file('protein.pdb')
prot.remove_all_waters()
prot.add_hydrogens()
for l in prot.ligands:
prot.remove_ligand(l.identifier)
For best results, manually check proteins before submitting them for calculation.
Once the protein is prepared, the hotspots.calculation.Runner object can be
used to perform the calculation:
from hotspots.calculation import Runner
runner = Runner()
# Only SuperStar jobs are parallelised (one job per processor). By default there are 3 jobs, when calculating charged interactions there are 5.
results = runner.from_protein(prot, nprocesses=3)
Alternatively, for a quick calculation, you can supply a PDB code and we will prepare the protein as described above:
runner = Runner()
results = runner.from_pdb("1hcl", nprocesses=3)
The hotspots.hs_io module handles the reading and writing of both hotspots.calculation.results
and hotspots.best_volume.Extractor objects. The output .grd files can become quite large,
but are highly compressible, therefore the results are written to a .zip archive by default,
along with a PyMOL run script to visualise the output.
from hotspots.hs_io import HotspotWriter
out_dir = "results/pdb1"
# Creates "results/pdb1/out.zip"
with HotspotWriter(out_dir) as writer:
writer.write(results)
If you want to revisit the results of a previous calculation, you can load the
out.zip archive directly into a hotspots.calculation.results instance:
from hotspots.hs_io import HotspotReader
results = HotspotReader('results/pdb1/out.zip').read()
While Fragment Hotspot Maps provide a useful visual guide, the grid-based data can be used in other SBDD analysis.
One example is scoring atoms of either proteins or small molecules.
This can be done as follows:
from ccdc.protein import Protein
from ccdc.io import MoleculeReader, MoleculeWriter
from hotspots.calculation import Runner
r = Runner()
prot = Protein.from_file("1hcl.pdb") # prepared protein
hs = r.from_protein(prot)
# score molecule
mol = MoleculeReader("mol.mol2")
scored_mol = hs.score(mol)
with MoleculeWriter("score_mol.mol2") as w:
w.write(scored_mol)
# score protein
scored_prot = hs.score(hs.prot)
with MoleculeWriter("scored_prot.mol2") as w:
w.write(scored_prot)
To learn about other ways you can use the Hotspots API please see the examples directory and read our API documentation.