Tutorial. Docking School SAnDReS Tutorial Cyclin-Dependent Kinases with K i Information (Introduction)

Transcription

1 Tutorial Docking School SAnDReS Tutorial Cyclin-Dependent Kinases with K i Information (Introduction) Prof. Dr. Walter Filgueira de Azevedo Jr. Laboratory of Computational Systems Biology azevedolab.net azevedolab.net walter@azevedolab.net 1

2 Introduction to SAnDReS 2

3 Introduction to SAnDReS SAnDReS is not a new docking program! 3

4 Introduction to SAnDReS SAnDReS is not a new boy band! 4

5 Introduction to SAnDReS What is SAnDReS? SAnDReS is a free and open-source (GNU General Public License) computational environment for the development of machine-learning models for prediction of ligandbinding affinity. SAnDReS is also a tool for statistical analysis of docking simulations and evaluation of the predictive performance of computational models developed to calculate binding affinity (Xavier et al., 2016). SAnDReS was developed in Python 3, using the SciPy, NumPy, scikit-learn, and Matplotlib libraries. 5

6 Introduction to SAnDReS There are different approaches to protein-ligand docking simulations, but if we consider selection of the biomolecular system, docking simulations and the validation phases, they are all included in the flowchart below. Independent of the programs employed in the study. Selection of Biomolecular System Re-Docking Statistical Analysis of Ensemble- Docking Results Virtual Screening with a Dataset with Decoys + Actives Statistical Analysis of Virtual Screening Simulation Results Download Structures and Experimental Binding Information Statistical Analysis of Redocking Results Scoring Function Calculation Statistical Analysis of Virtual Screening with Decoys + Actives Best Hits Pre-Docking Analysis Ensemble- Docking Statistical Analysis of Scoring Functions Results Virtual Screening Simulation 6

7 Introduction to SAnDReS Our goal in developing SAnDReS is to have an integrated tool for development of machine-learning models and statistical analysis of docking results in order to improve the docking accuracy. In addition, we also developed tools to analyze the influence of structural parameters on docking results. Most of these structural parameters are based on crystallographic information available in the PDB file for a given structure. The decision to focus on the crystallographic information is due to the fact that the majority of the structural information available for protein in complex with ligands for which there is experimental binding information is derived from x-ray crystallography technique. As we can see in the table below, over 94 % of structures of complexes, for which there is binding information, were obtained by X-ray crystallography. Binding Affinity Total Number of Total Number of Total Number of Total Number of Total Number of Total Number of Available Structures Structures Determined Structures Structures Structures Structures by X-ray Determined by Determined by Determined by Determined by Crystallography NMR Neutron Crystallography hybrid methods electronic micrography Ki Kd Ka ,0 IC DeltaG DeltaH Percentage of Structures Determined by X-ray Crystallography Search carried out on December 8th

8 Introduction to SAnDReS The flowchart below brings the SAnDReS participation in a docking project. The green boxes can be carried out by SAnDReS. Selection of Biomolecular System Re-Docking Statistical Analysis of Ensemble- Docking Results and Structural Parameters Statistical Analysis of Scoring Functions Results Virtual Screening Simulation Download Structures and Experimental Binding Information Statistical Analysis of Re-Docking Results Scoring Function Calculation Virtual Screening with a Dataset with Decoys + Actives Statistical Analysis of Virtual Screening Simulation Results Pre-Docking Analysis Ensemble-Docking Building New Scoring Functions Results Using Machine Learning Statistical Analysis of Virtual Screening with Decoys + Actives Best Hits 8

9 Introduction to Tutorial 001 In this tutorial, we carry out a full analysis of docking results and also a study involving the development of machine-learning models for cyclin-dependent kinase (CDK) (EC # ). This enzyme has been studied as a protein target, mainly because its role in controlling cell-cycle progression (Morgan, 1995; Murray, 1994). For recent reviews see: De Azevedo, 2016; Levin et al., 2016) Specifically in this tutorial, we are going to study crystallographic structures of CDK in complex with inhibitors for which there is experimental information for binding affinity. We work with an ensemble of CDK structures for which there is information for Ki. (De Azevedo et al., 1997). EC = Enzyme classification 9

10 Running on the Windows 10

11 Installing Windows Version You need to have Python 3 installed on your computer to run SAnDReS. In addition, you also need NumPy, Matplotlib, scikit-learn, and SciPy. You can make the installation process easier by installing pyzo. While installing them, let's the music play... Step 1. Install the Pyzo IDE (download here) Step 2. Install Python environment (download here) Step 3. Install Scientific packages needed to run SAnDReS. Open Pyzo IEP. To run Pyzo IEP, go to c:\program Files (x86)\pyzo directory and you will have the pyzo IEP. Double click on pyzo. In the Pyzo s shell (IEP), type the following commands: conda install numpy conda install scipy matplotlib conda install scikit-learn Don t forget to press <Return> after typying each line. 11

12 Installing Windows Version Step 4. Download SAnDReS (download here) Step 5. Unzip the zipped file (sandres.zip) Step 6. Copy sandres directory to c:\. Step 7. Open a command prompt window and type: cd c:\sandres python sandres1_gui.py Don t forget to press <Return> after typying each line. This launches GUI window for SAnDReS. That s it, good sandres session. You can also start SAnDReS clicking on the sandres.bat file ( ). You may also create a shortcut for SAnDReS right clicking on the sandres.bat file. 12

13 Running in the Windows When we start a new SAnDReS session, the program checks if everything is fine and launches GUI window. The window below, called text window, is kept open during a SAnDReS session. 13

14 Running in the Windows GUI window Text window 14

15 Running in the Windows Project directory entry From now on we can use the GUI window to work with SAnDReS. In the project directory entry, we have to browse to find a pre-existing directory, where all work will be carried out. 15

16 Running in the Windows SAnDReS GUI Input file (.in) SAnDReS main program log files (.log) csv files (.csv) Plot files Machine-learning techniques SAnDReS GUI is a front-end that calls SAnDReS main program, that carries out most of the computing work. In addition, SAnDReS has a set of machinelearning techniques implemented in the code scikit_regression_methods_v1_1. The GUI takes care of preparing input files that will be run by SAnDReS main program. SAnDReS was developed in a way that you can perform all its tasks from the GUI window. You can also edit the input file and run it, using the Run It green button on the SAnDReS GUI window. 16

17 Running in the Windows Below we have a description of the GUI window. Update button for project directory Menu commands Browse button for project directory Clear button for input file name Run button (to run a specific input file) Log screen Import button for re-docking file name Import button for ensembledocking file name Information strip Clear button for re-docking file name Clear button for ensemble-docking file name Clear button for score file name Import button for score file name Exit button Clear Log screen button 17

18 Running in the Windows Below we have a description of the entries in the GUI window. Input file name, to indicate an input file to be run when we click on the Run It button. Entry for ensemble-docking file name Entry for project directory. For each docking project you should create a directory, where all structures, CSV files, input files, log files, and plots will be stored. Every time you start a new docking project, you browse a pre-existing directory. Entry for re-docking file name Entry for scoring function file name The user can change the file names, the names shown here are the default ones. They will be the ones used in the tutorial

19 Copying Tutorial 001 Files 19

20 Copying Tutorial 001 Files Now we are going to copy all files necessary for the tutorial 001. This tutorial highlights major features of SAnDReS. The biological system to be analyzed in the tutorial 001 is cyclindependent kinase, for which there is experimental information for inhibition constant Ki, and, of course, crystallographic structures. Consider that you created the C:\Users\Walter\Desktop\T 001 directory. Click on the Find button and then you have to navigate to where the directory is located. 20

21 Copying Tutorial 001 Files Click on the selected directory. SAnDReS takes few seconds to update the project directory. 21

22 Copying Tutorial 001 Files Our project directory is C:\Users\Walter\Desktop\T 001\ You may copy and paste the directory information into the Project Directory field. Don t forget to end with \. 22

23 Copying Tutorial 001 Files One important feature the users should be aware of is the pre-defined preferences in SAnDReS. To modify them click on File->Preferences, as shown here. 23

24 Copying Tutorial 001 Files On the new pop-up window, we have the URLs used by SAnDReS to download files from the Protein Data Bank. Since the PDB is in continuing development, sometimes the PDB developers change the directories. The users can update this information using this menu. Another important information is related to Google Chrome directory. That should be informed if the users want to access online tutorials through SAnDReS. 24

25 Copying Tutorial 001 Files The users can also change the radius of the binding sites, which are generated during the Pre-Docking phase. Yet another input file used by SAnDReS is related with ligand codes normally used in crystallization conditions. These ligands are omitted from the structures in the dataset. Click on the Close button. 25

26 Copying Tutorial 001 Files To copy tutorial files to project directory, click on File->Copy tutorail001 to Project Directory, as shown here. 26

27 Copying Tutorial 001 Files Now we have a new popup window with a summary for this tutorial. To copy the files needed to run this tutorial, click on the Copy Files button. 27

28 Copying Tutorial 001 Files Information about all files copied to the Project directory is shown in the Log screen. We may now click on the Close button on the Tutorial pop-up window. 28