Tutorial

This is the tutorial page for the MORLD web service.
Follow the tutorial step by step.

• Demo
  
• Queue Page
  
• Result Files
  

Demo

---

In this tutorial, we will optimize the compound, ponatinib, to generate novel inhibitors against the target protein, DD1 kinase. You can download the prepared pdb file of DDR1 kinase here.

According to Autodock Vina, Ponatinib shows the binding energy around -13.1 kcal/mol against DDR1. Using MORLD, We will decrease the binding energy of the initial molecule, ponatinib, against DDR1.

If you want to check the result of demo quickly, please refer to the JobID #1 at Queue page.

---

• Step 1
• To submit the job, go to "Job Submission" page.

---

• Step 2
• Insert the job name. For example, demo_pona

• (Optional) If you want to be notified of job termination, enter your email at "Email" field.

• (Optional) If you want to secure your job, enter the password at "Password" field. The result file will be secured by 7z format with the password.

---

• Step 3

• Insert the canonical SMILES string of ponatinib at the text field of "Initial molecule".

  CC1=C(C=C(C=C1)C(=O)NC2=CC(=C(C=C2)CN3CCN(CC3)C)C(F)(F)F)C#CC4=CN=C5N4N=CC=C5

---

• Step 4
• Upload the prepared pdb file of the target protein, DDR1 3zosA_prepared.pdb.

• To prepare a PDB file for MORLD, remove all water molecules, and ligands in the PDB structure. Also, the protein needs protonation for accurate docking. Use protonation programs like PDB2PQR server.

---

• Step 5
• (Optional) You can modify the parameter of MORLD model at "Modification Configuration". You can pass this step with the default setting.

• Atom types: You can customize the atom types for modification. The atom types should be the atomic symbols as like ones in SMILES string. Select the atom types by check boxes or enter your own atom types with "," delimiter. A default value is set to C, O, N.

• Max num. of modification: MORLD will add or remove atoms by this given number. You can set the number of modifying action at each episode. The larger value will allow to modify the initial molecule more differently. On the other hand, small value will produce similar structure with the inital one.

• Num. of training episode: MORLD will generate optimized molecules by this given number. *If an error occurs during the docking process, the number of molecules generated may be small than this number.The higher the number, the more experience of modification that MORLD gains and, consequently, the more likely to get better results. Usually, at least 5000 episodes are need to get moderate results.

• Weight for SA score: You can customize a weight value for synthetic accessibility (SA) score. The default value is set to 1 but you can change to 0 (not consider SA score) or larger than 1 (more focus on SA score).

• Weight for QED score: You can customize a weight value for quntitative estimate of drug-likeness (QED) score. The default value is set to 1 but you can change to 0 (not consider QED score) or larger than 1 (more focus on QED score).

---

• Step 6

• Insert the binding site information. The given values are derived from PDB ID:3ZOS and ADT (AutoDockTools).

  Center x: -7.5
  Center y: 2.5
  Center z: -40
  Size x: 24
  Size y: 20
  Size z: 20

---

• Step 7

• Click the "Submit" button and check the job has been submitted at "Queue" page.

  With the default setting, it will takes about two days. If you want to run a test trial, reduce the parameter "Max num. of modification" or "Num. of training episode".

Queue Page

---

Check the submitted job at "Queue" page.

---

• Job ID
• ID number of the submitted job.

---

• Status

• Not starged yet: The job is submited to the server but not started yet. If the job is not started within few minutes, please check the input information again or contact to the server manager (jjs092@kaist.ac.kr).

  Running: The job is successfully running on the server.

  Finished: The job is finished and you can download the result file.

  Error occurred: The job is stopped by error. Please check the input information whether it is entered correctly. For more information on the error, we provide "error_log.txt" file, the error log of the job. You can download the log file via download link of the queue page.

---

• Job
  Name
• The name of job submitted.

---

• Submitted
  Time
• Submitted time of the job.

---

• Donwload
  Link

• If the job has finished successfully, then the Download link of the result will appear.
  
  If the job has occured an error, then an "Error" link will be appread. The "Error" link contains detailed information about the error.

Result Files

---

Inside the download file, you will find below files and folder.

---

• optimized_
  total.csv

• This files contains all the optimized molecules. The number of optimized molecules are the same as the number of training episode. If you set the number of training episode as 5000, then there are 5000 optimized molecules in "optimized_total.txt" file.

  Each column represents the generated number, SMILES code of the generated molecule, binding energy (Qvina score), SA score, and QED score respectively. With this data, you can sort by the binding energy to find the highest binding affinity molecule. Or set the minimum SA score to filter the irrational structure of the optimized molecules.

---

• log_
  run.txt

• This file is a log of the MORLD model. It shows in what order the molecules are modified at each Episode.

---

• docking_
  pose/

• This folder contains the files of binding pose of the optimized molecule. Each file is named as "pose_##.pdbqt" where ## is the generated number. You can see the binding pose of the highest docking score molecule in "optimized_total.txt" file by checking the generated number.

  You can open .pdbqt file and see the binding pose with the target protein with visulization tool such as Pymol.

---

• (job ID)_
  info.txt

• This file is contains the input paremeters of the job.