sb-ncbr/confanalyser

ConfAnalyser

Import ConfAnalyser and run from another python script:

from ConfAnalyser import ConfAnalyser

result = ConfAnalyser(paths_file=paths_to_pdbs_path, names_file=atom_names_path,
             molecule_type=ConfAnalyser.MoleculeType.Cyclohexane).result()

Run as a command-line application: python3 ConfAnalyser.py -l paths_to_pdbs -n atom_names --(cyclohexane|cyclopentane|benzene)

Electron density coverage analysis scripts

Installation

No installation is required. After cloning the repository, change directory to electron_density_coverage_analysis.

Requirements

Python version 3.10 or higher gemmi package versions 0.6.5, can be installed using pip, e.g.:

pip install gemmi==0.6.5

Running analysis for a single structure

Use electron_density_coverage_analysis.py, which is a program that determines electron density coverage of a cycle. By default, uses trilinear interpolation to infer the electron density value, and considers an atom to be covered by the electron density when the corresponding intensity is MORE than the threshold for the isosurface (1.5 sigma).

Arguments description

Positional arguments (required)

• input_cycle_pdb Input PDB file with coordinates of a cycle, produced by PatternQuery
• input_density_ccp4 Input electron density file for the corresponding protein structure from the PDB in CCP4 format

Mandatory arguments

Choose one of the two modes:

• -s Simple mode - output is two numbers: first is the number of covered atoms, the second is the total number of atoms in a cycle
• -d Detailed mode - output a sequence of paried values, where the first member of each pair is a serial atom number, and the second member is "y" - if atom is covered, and "n" - if it is not.

Optional arguments

• -m, --more_or_equal Atom is considered to be covered by the electron density when the corresponding intensity is MORE OR EQUAL to the threshold for the isosurface
• -c, --closest_voxel Instead of trilinear interpolation, the intensity of the closest voxel is used

Example of running analysis for a single structure

On Windows, we recommend to use Anaconda PowerShell Prompt

From electron_density_coverage_analysis directory run the script in e.g. simple mode with e.g. --more_or_equal setup:

python electron_density_coverage_analysis.py -sm tests/example_input/single_structure/3biu_0.pdb tests/example_input/single_structure/3biu.ccp4

Running analysis for multiple structures

It is possible to run the analysis for multiple structures using main.py script. By default, analysis uses trilinear interpolation to infer the electron density value, and considers an atom to be covered by the electron density when the corresponding intensity is MORE than the threshold for the isosurface (1.5 sigma).

Input should follow a specific format:

The names of directories should be always as on the picture (validation_data as a main folder, then folders for three cycle types, folder filtered_ligands in each one, then folders for each ligand, then patterns folder in each of them), which contains .pdb files of cycles produced by Pattern Query. Filename format should always be as in the example: {ligandID}_{pdbID}_{i}, ligandID is ligand ID can be up to 3 characters long.

Arguments description

Positional arguments (required)

• rootdir Root directory (e.g. "validation_data") with files containing coordinates of cycles, produced by PatternQuery. Directory hierarchy should follow a specific format (see above)

Optional arguments

• -m, --more_or_equal Atom is considered to be covered by the electron density when the corresponding intensity is MORE OR EQUAL to the threshold for the isosurface
• -c, --closest_voxel Instead of trilinear interpolation, the intensity of the closest voxel is used

Example of running analysis for a multiple structures

On Windows, we recommend to use Anaconda PowerShell Prompt

Create ./ccp4 folder in electron_density_coverage_analysis folder:

mkdir ccp4

Download .ccp4 files for each of the ligands in input folder and put them into ./ccp4 folder. E.g. for tests/example_input/validation_data input, you should download the following .ccp4 files:

• 4omc
• 3v8d
• 8gdi
• 8gxh
• 8gxi
• 8gxp

From electron_density_coverage_analysis directory run e.g.:

python main.py tests/example_input/validation_data -m

This will run the analysis in --more_or_equal mode.

Output of the script will be collected in ./output folder including:

• Three .csv files for each ligand type, e.g. cyclopentane_params_m_analysis_output.csv, in names of which parameters (m in this case) for analysis are specified. Each .csv file follow a specific format, e.g.:

06L_8gxp_0,06L,2;5
0GV_3v8d_1,0GV,5;5
0GV_8gdi_0,0GV,5;5
0GV_3v8d_0,0GV,5;5

where first column specifies ligand PDB file name without extension, second column specifies ligand, third column specifies coverage - first number is a number of covered atoms, then semicolon, then the total number of atoms in a cycle