Identification of Inhibitors against Hemagglutinin from the H5N1 Influenza Virus

1
Identification of Inhibitors against
Hemagglutinin from the H5N1 Influenza Virus

• The Ensemble based Virtual Screening approach
  with Relaxed Complex Scheme
• Kevin Wu
• CNIC, Beijing, China
• August 16, 2008

2
Specific Aims

• 1. RMSD cluster analysis of H5 using ptraj to
  generate representative snapshots of the H5
  hemagglutinins receptor binding domain from
  already compiled MD trajectories using NAMD
• 2. Using these representative structures, perform
  virtual screening using an Autodock 4 web service
  (ws.nbcr.net) to screen against the NCI diversity
  set to identify top hits for further evaluation.
  The Virtual Screening results will be loaded into
  the Avian Influenza Virus virtual screening
  database developed and deployed at CNIC.
• 3. Using the relaxed complex scheme (RCS), the
  top hits identified from virtual screening of the
  NCIDS will be redocked to all of the H5 cluster
  representatives. These results will be rescored
  using the weighted harmonic means method in order
  to generate a more comprehensive and reliable set
  of results.

3
Current Progress

• Since my last report, I have begun to experiment
  with docking disaccharides through Autodock
• To help save myself the trouble manually editing
  each docking parameter file to vary the
  parameters that allow for an accurate redocking.
  I wrote a bash script that loops from an user
  determined start to end value by an increment
• By using this script that I wrote, this allows me
  to vary only one parameter, so that I can
  accurately analyze the improvement of increasing
  and decreasing each parameter
• One of the features I have added is that after
  docking, it will generate both the conformation
  of the lowest energy ligand in the largest
  cluster as well as all of the conformations of
  the ligands in the docking log file. It will then
  calculate RMSD of each conformation against the
  original crystal pdb ligand file and write the
  values to separate text files which are then
  sorted by RMSD value

4
Results

• By using this convenient script Ive written,
  Ive been able to test hundreds of redockings
  with different variations and quickly evaluating
  results via RMSD. However, I recently found that
  by using RMSD values of only the largest
  clusters lowest energy ligand, I disregard
  results of other potentially useful clusters
• With the data Ive gathered from using only the
  largest clusters lowest energy, most of it does
  not make much sense to me, further evaluations
  and analysis will be needed. The RMSD values will
  make more sense with further analysis of the dlg
  files by visualization in autodocktools
• However, I have successfully redocked sialic acid
  back into a hemagglutinin receptor (PDB ID 1jsn)
  with a relatively low RMSD (lt 2.0).

5
Next Week

• Improvement on the script is still being made to
  further better organization of the generated
  information. Also, adding functionality to
  generate all RMSD value of all clusters will
  still need to be implemented
• Excel graphs will be made of the RMSD values to
  visualize them and to try and make further sense
  of the data in order to try to change the correct
  parameters to allow for redocking of
  disaccharides and possibly trisaccharides
• Upon correctly redocking of the sialic acid-gal
  disaccharide, I will attempt virtual screening of
  the 1jsn crystal structure with the NCI diversity
  set

6
Successes!

• One of my biggest successes has been thinking up
  the script that allows multiple submissions with
  with different parameters
• If this script had not been written up, I could
  have spent hours submitting jobs manually
• However, this was also due to the fact that I had
  read extensively about shell scripting and sed
  which led me to this idea
• One of the current successes that I have had is
  redocking sialic acid back into the hemagglutinin
  receptor 1jsn

7
Cultural Experiences
Top Left Beach Volleyball,
USA vs. Switzerland Top Right American fans at
Beach Volleyball game Middle
Right Qianmen gate at night Middle Left Eating
Peking Duck at Da Dong Bottom Lotus at Old
Summer Palace
8
Acknowledgements

• Dr. Wilfred Li, Dr. Dong Xu, Dr. Rommie Amaro,
  Lily Cheng, and Hsing Pao for helping me fully
  understand my project and for giving me direction
• Dr. Nan Kai, Dr. Zhonghua Lu, and Dr. Jianjun Yu
  for hosting and mentoring me at CNIC
• Guangyuan Lu and Yanhua Sun for help with the
  various computing and scripting problems Ive had
• Haiyan Xu and Wei Chen for arranging PRIME at
  CNIC
• Dr. Gabriele Wienhausen, Dr. Peter Arzberger,
  Teri Simas, and Tricia Taylor Oliveira for
  allowing PRIME to happen
• The NSF, Calit2, and PRIME for this wonderful
  opportunity!
• Michael Wang for the great experiences and
  companionship