TASSFUN - Monthly Report

August 2004

• starting at EML Research
• making the PhD-exam (3 days in Munich)
• getting to know the infrastructure of EML (ID card, computer account, telefon,...)
• getting on overview of the new literature
• checking protonation state and hydrogens of given pdb files
• converting pdb files for using in amber (C++ programm)
• creating a concept for putting the data into mysql
• installation of mysql on priamos (together with Jan)

September 2004

• putting programs for converting pdb files into bash-script for running automatic procedure 
• problems with assigning atom types of ligand with antechamber
• setting new parameters for lead component of the investigated ligand.
• trying different restraints for minimization in sander
• figure for a review

October 2004

• figure for a review (pdf, 4.6 MB)
• first Combine analysis
• 3-months-report (28.10.2004)
• telephone conference (28.10.2004)
  

November 2004

• conference at EMBL "Structures in Biology"
• GDCh - 18. CIC-Workshop in Boppard
• preparing data set of structures of uPA for COMBINE anlaysis
• starting cooperation with Olga Krebs (EML Research, Scientific Databases and Visualization) for designing of database for protein-ligand interaction complexes (first concept)
  
• removing some bugs of the programs for preparing pdb files

December 2004

• writing program for calculating interactions with anal (amber7)
  
• preparing data set of structures of uPA for COMBINE anlaysis
• learning GOLPE
  
• Combine analysis on 31 uPA structures (uro1)
  
• modification of programs for reading various pdb-files without hydrogens
• problem of adding hydrogens to small molecules
• christmas holidays

January 2005

• implementation of adding hydrogens to the protein and ligand by WHATIF and insightII, respectively
• problems of assigning atom types with antechamber
• 6-months-report

February 2005

• building new urokinase model
  
• preparing visit at AstraZeneca
• visit at AstraZeneca (19.-25. Febr. 05)
• problems of assigning parameters with antechamber are solved by using SDF files
• bug fix in some of the scripts and programs
  

March 2005

• Creating SDF files with all important entries (ID, residue, netcharge, PDB).
• Compare minimizations of 1o5a (1o5a.min.tar.gz)and 1o5c (1o5c.min.tar.gz) with different protonation states and restrains. For next COMBINE models structures with HID57 and phenolOH will be used.
  
<>36 structures of urokinase were aligned with pymol to structure 1o5a.pdb (uPA.align.pdb.tar.gz) and the different sidechain conformations in the active site were compared (superposition of different uPA structures, compare_target.tif). Three important alternative conformations could be found (His99, Gln192, Arg217).

general template: 1o5a
His99: A (not uniform conformation), template: 1sqt; B (more likely than A), template: 1o3p
Gln192: A, template: 1c5z; B (prossibly mixture of conformation A and C), template: 1gi8; C, template 1o3p
Arg217: A (not or very week defined in the available electron density maps), template: 1c5z; B (well defined), template: 1owd

Models were changed:
ALA190 changed to SER190 with conformation from 1o3p.



1o5a
model (uPA-models.tar.gz)	His99	Gln192	Arg217
1 (model1.pdb)	A (1sqt)	A (1c5z)	B (1owd)
2 (model2.pdb)	B (1o3p)	A (1c5z)	B (1owd)
3 (model3.pdb)	A (1sqt)	B (1gi8)	B (1owd)
4 (model4.pdb)	B (1o3p)	B (1gi8)	B (1owd)
5 (model5.pdb)	A (1sqt)	C (1o3p)	B (1owd)
6 (model6.pdb)	B (1o3p)	C (1o3p)	B (1owd)

The six models were in sander by 50 steps (restrain: mainchain 500, sidechain 50, residue H99/Q192/R217 1).
(uPA xray structures and their alternative conformations)
selected water molecules


• Change the programs convert.sh and xleap.sh for running via command line.
• Table with residue names, netcharges, Ki,... (table)
  

April 2005

• Selection of important water molecules for ligand binding in uPA.
• Assigning parameters with antechamber.
• 9-month-report
• Poster for "19th Darmstadt Molecular Modelling Workshop in Erlangen"

May 2005

• "19th Darmstadt Molecular Modelling Workshop in Erlangen" (3.-4.5.2005) (poster)
  
• preparation of files for visit in Mölndal
  

June 2005

• vacation 26th May - 19th June
  
• visit in Mölndal 31st May - 1st June
• docking of ligand of 1ghw into active site of model1
  
• Cluster Analysis of water sites of urokinase-ligand-complexes
• EML lab report
  

July 2005

• adapting changes made in Mölndal to the programs and scripts
  
• testing various minimization protocols with different dielectric constants (docking)
• learning more about GOLD and other docking methods
• learning CORINA
  
• paper review

August 2005

• preparation and correction of SD file (all_300805_3D.sdf, PDB_uPA_300805_3D.sdf)
  
• docking experiments in GOLD (PDF) (still in process!)
  

September 2005

• visualization of molecular interactions (red <=-1, white, blue >1; model5-elec-surface.png, model5-elec-surface-ligands.png, model5-vdw-surface.png, model5-vdw-surface-ligands.png, model5-vdw.avi, model5-elec.avi)
  
• new COMBINE models
• new COMBINE model with strong constraints on minimized training set
• Results of new COMBINE models

October 2005

• various COMBINE models for urokinase together with docking results
• 15-month-report (word)

November 2005

• optimizing UHBD protocol
• building urokinase COMBINE models with solvation energy terms (problems with outliers)
  
• correcting Ki values (all_301105_3D.sdf.gz)
• building new urokinase COMBINE models with changed Ki values and testing for outliers
• analysing UHBD solvation energy terms
• urokinase COMBINE model (final model html, varous models excel): various models tested with re-docked ligands (with changed Ki values and solvation energy terms)
  
• annual report
  
• 1^st German Conference on Chemoinformatics / 19. CIC-Workshop (13. - 15. November 2005, Goslar)
• Medientraining
  

December 2005

• German Biotech-Industrie goes Wertschöpfung (1.12.2005, Industriepark Höchst)
• new trypsin model
• preparation of a thrombin model
• SD file
• docking ligands of Boehm, Stuerzebecher, Klebe into bovine trypsin model
• re-docking of ligands into urokinase (flood fill 20 A), docking solutions: best_ranked.sdf ranked.sdf

January 2006

• problems with large disk space requirements for jobs with several hundreds of docking solutions
• parallelistation of scripts for docking/minimization/calculation for running on a compute-cluster
  
• re-docking of ligands into urokinase receptor model (flood fill 20 A) and prediciton of dG by uPA COMBINE model
  ranked and rescaled predicted dG values (excel file)
  
• docking of ligands with known Ki but unknown x-ray conformation and prediction of dG by uPA COMBINE model
  testset-rankedA, testset-rescaledA, testset-rankedP, testset-rankedQ
• re-docking of ligands into trypsin receptor model and prediction of dG by trypsin model
• docking of ligands of Boehm, Stuerzebecher, and Klebe (J. Med. Chem. 1999) into trypsin model and prediction of dG
  
• preparation of structures for thrombin model
• quarter report
  

February 2006

• preparation of structures for thrombin model
• visualization of interaction energies and partial weight coefficients (figures and details for annual report)
• summary of results of docking experiments (uPA, trypsin, thrombin)
  

March 2006

• vacation February 25th - March 4th
  
• visit in Mölndal (March 6th - 10th) for installing scripts and programs (summary doc html)
  
• docking of trypsin-like protease binders and nonbinders into urokinase model
• learning of JMP6 for showing results of docking
  
• preparation poster for Keystone Symposia
  

April 2006

• Keystone Symposia: Structure Based Drug Discovery (April 4th - 9th) poster
• Docking of ligands of training set and test set (with available X-ray structures)
• docking-1c4u-PDB_thrombin-190406 with PDB coordinates and RMSD
  thrombin model2 1c4u with sodium:
  /home/henricsn/combine2go/data/thrombin/model2_180406/anal/output-A91.dat
  ChemScore, 28 structures, 592 variables
• Changing thrombin model from apo-structure to x-ray structure 1c4u without sodium
  /home/henricsn/combine2go/data/thrombin/model2_260406/anal/
• analysing conserved water sites of thrombin structures
  

May 2006

• Heidelberger Innovationsforum (8.-9. May)
• thrombin model 3 (1c4u but with rotated lysine) and docking of ligands
  

June 2006

• new trypsin model with 48 ligands
  
• docking of ligands into trypsin model for publication
• docking of ligands into thrombin model for publication
• model validation: creating different subsets (thrombin model 3, thrombin model 4)
  
• statistical analysis of predicted values (thrombin)
  
• RMSD calculation for docked ligands
• try to find a suitable filter by selecting best ranked values, eg. Gold.Goldscore.Fitness, lowest predicted dG, dG_bind_elec, sum_elec
  
• preparation of publication
  

July 2006

• changing and recalculation of COMBINE model of urokinase for publication
• statistical analysis of predicted values
• preparation of publication (histograms, ROC and REC curves)
• final summary (row data, figures)
  

August 2006

• preparation of publication
• holiday
  

September 2006

• MMS06: Methods of Molecular Simulation 2006
• building a COMBINE model of thrombin with four conserved water molecules in the active-site
• preparation of publication
• paper review
  

Oktober 2006

• Workshop on Molecular Targets for Cancer, Luxembourg, 6.-7.10.2006
• kick-off meeting, Modena 16.-19.10.2006
• preparing a poster for Goslar
• analyzing trypsin ligands
  
• Zn²⁺ mediated ligand binding in thrombin
  

November 2006

• 2nd German Conference on Chemoinformatics, Goslar, 12.-14.11.2006
  
• Accelrys, European Science Forum, Heidelberg, 28.11.2006 (Presentations, Posters)
• analysis of ligands of Howard et al, JMC 2006, 49, 1346 (new images)
  
• analysis of the correlation between LV 1 and 2 of a PCA of the training with the values retrieved from a PCA prediction for docked ligands
  
• STAPAC

December 2006

• Ligand interaction with the S1sub pocket of urokinase
  

March 2007

• low PLS coefficients with high energy terms