Structure Filtering Tutorial
=============================

This tutorial aims to guide the user to perform a database filtering
with the possibility of specifying which R-groups we want fixed.

Flags and Input Files
~~~~~~~~~~~~~~~~~~~~~

-i TEMPLATE_LIGAND, –template_ligand –> Path to PDB template ligand. 

-l LIGANDS, –ligands –> Path to SDF file with database ligands or folder with SDF files. 

-o OUTFILE, –outfile –> Output file name. 

-a ATOM_LINKER, –atom_linker –> PDB atom name of core that is bound to R-group.


Requirements of the input files
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1. The file of the substructure we want to have in all the molecules,
   must be in PDB format and have **unique PDB atom names**.
2. The database files must be SDF format.

Example - Running in a Jupyter Notebook
---------------------------------------

.. code:: ipython3

    import database_filtering
    from database_filtering.utils.utils import filter_mols
    import rdkit
    from rdkit import Chem
    from rdkit.Chem import AllChem
    from rdkit.Chem import Draw
    from rdkit.Chem.Draw import rdMolDraw2D
    from rdkit.Chem.Draw import IPythonConsole
    IPythonConsole.molSize = 300,300

.. code:: ipython3

    template_ligand = Chem.MolFromSmiles("O=C(O)c1ccccc1")
    template_ligand




.. figure:: ../../_images/DF_tutorial-input_ligand.png
    :scale: 40%
    :align: center



.. code:: ipython3

    ligands = [Chem.MolFromSmiles("O=C(O)c1ccccc1CO"), 
               Chem.MolFromSmiles("Nc1ccccc1C(=O)O"), 
               Chem.MolFromSmiles("Nc1ccccc1C(=O)O"),
               Chem.MolFromSmiles("O=C(O)c1ccccc1I"),
               Chem.MolFromSmiles("O=C(O)c1ccccc1Br"),
               Chem.MolFromSmiles("NC(=O)c1ccccc1C(=O)O"),
               Chem.MolFromSmiles("O=[IH2]c1ccccc1C(=O)O"),
               Chem.MolFromSmiles("Nc1ncccc1C(=O)O"),
               Chem.MolFromSmiles("Nc1ccc(C(=O)O)c(O)c1")]
    Draw.MolsToGridImage(ligands,molsPerRow=4,subImgSize=(200,200)) 





.. figure:: ../../_images/DF_tutorial-ligands.png
    :scale: 40%
    :align: center



.. code:: ipython3

    linker = ['C7']
    template_ligand.__sssAtoms = [8] # Highlight the atom C7
    template_ligand


.. parsed-literal::

    REMEMBER: If your linker is atom C7, we will only obtain molecules that have an R-group bound to that atom.
    Also, we can select more than one linker atom:
    In a notebook --> linker = ["C7", "C4"]
    Running on the cluster --> -a C7 C8 




.. figure:: ../../_images/DF_tutorial-linker_atom.png
    :scale: 40%
    :align: center



.. code:: ipython3

    template_ligand_path = "./template_ligand.pdb"
    ligands_path = "./ligands.sdf"

.. code:: ipython3

    filter_mols(template_ligand_path, ligands_path,'test',linker)
    # Results are stored in the file test.sdf 


.. parsed-literal::

    Filtering passed for molecule 
    Filtering passed for molecule 
    Filtering passed for molecule 
    Filtering passed for molecule 
    Filtering passed for molecule 
    Filtering passed for molecule 
    Filtering passed for molecule 
    No substructure match for ligand , skipping
    Molecule  did not meet the R-groups requirements.


Filtering Results
~~~~~~~~~~~~~~~~~

.. code:: ipython3

    Draw.MolsToGridImage(mols,molsPerRow=4,subImgSize=(200,200)) 




.. figure:: ../../_images/DF_tutorial-results.png
    :scale: 40%
    :align: center



Running in the Cluster
----------------------

.. code:: sh

   #!/bin/bash
   #SBATCH -J filter
   #SBATCH --output=filter.out
   #SBATCH --error=filter.err
   #SBATCH --ntasks=3
   #SBATCH --mem-per-cpu=10000

   source /shared/home/hmartin/miniconda3/etc/profile.d/conda.sh
   conda activate /shared/home/hmartin/miniconda3/envs/r_groups_env

   python -m database_filtering.run_filtering -i template_ligand.pdb -l ligands.sdf -a C7 -o test