AI helps you reading Science

AI generates interpretation videos

AI extracts and analyses the key points of the paper to generate videos automatically


pub
Go Generating

AI Traceability

AI parses the academic lineage of this thesis


Master Reading Tree
Generate MRT

AI Insight

AI extracts a summary of this paper


Weibo:
The parameters which control the nature of the conformational ensembles were varied in a statistical manner in order to optimize the performance of OMEGA for the purposes of reproducing bioactive conformations of protein-bound ligands

Assessing the performance of OMEGA with respect to retrieving bioactive conformations.

Journal of Molecular Graphics and Modelling, no. 5 (2003): 449-462

Cited by: 304|Views7
WOS EI

Abstract

OMEGA is a rule-based program which rapidly generates conformational ensembles of small molecules. We have varied the parameters which control the nature of the ensembles generated by OMEGA in a statistical fashion (D-optimal) with the aim of increasing the probability of generating bioactive conformations. Thirty-six drug-like ligands fr...More

Code:

Data:

0
Introduction
  • Searching databases with the aim of finding structures that are similar in some fashion to a given query structure is of great interest in ligand-based design
  • Such searches are typically based on either molecular graph techniques (2D) [1] or superposition (3D) methods [2].
  • The great advantage of using 3D over 2D methods is that they explicitly take shape-dependent properties into account
  • In this way, 3D methods increase significantly the probability of finding structures which have similar shape and chemical properties to the query structure, but which are less intuitively related according to chemical class.
  • A pre-requisite for obtaining reliable results when using pre-calculated conformations (e.g. OMEGA and Catalyst) [5], is that representations of the ligands bioactive conformation must be present in the multi-conformer database
Highlights
  • Searching databases with the aim of finding structures that are similar in some fashion to a given query structure is of great interest in ligand-based design
  • At 1.0 Å the hit-rate doubles when the input structure to OMEGA is from Corina, whereas importantly there is almost no change in the hit-rate when the pre-optimization step is carried out
  • The parameters which control the nature of the conformational ensembles were varied in a statistical manner in order to optimize the performance of OMEGA for the purposes of reproducing bioactive conformations of protein-bound ligands
  • Thirty-six drug-like ligands determined by high-resolution X-ray crystallography have been analyzed
  • Molecules with few torsional degrees of freedom were found to be less sensitive to the various settings, whereas the bioactive conformations of structures having eight or more rotatable bonds proved difficult to retrieve
  • The majority of the structures were found to bind in low-energy conformations, in particular when using MMFF94s to generate input structures
Methods
  • OMEGA uses a depth-first searching algorithm for generating conformational ensembles.
  • It is a rule-based method that generates conformations extremely rapidly.
  • OMEGA disassembles the molecule into fragments of rotatable bonds, and reassembles the fragments based on the sorted order of the fragment energies—the depth-first method.
  • OMEGA uses a modified version of the Dreiding force field, which does not include any electrostatic terms.
  • Note that the input bond lengths, bond angles
Results
  • At 1.0 Å the hit-rate doubles when the input structure to OMEGA is from Corina, whereas importantly there is almost no change in the hit-rate when the pre-optimization step is carried out.
  • This graph makes it possible to compare the hit-rate to other studies, in which a different criterion to specify a match to a bioactive conformation is chosen
Conclusion
  • The parameters which control the nature of the conformational ensembles were varied in a statistical manner in order to optimize the performance of OMEGA for the purposes of reproducing bioactive conformations of protein-bound ligands.
  • The authors recommend setting the GP ENERGY WINDOW parameter to a low value (≤5 kcal/mol), the GP RMS CUTOFF parameter to a low value (≤0.6 Å), and generating as large conformational ensembles as feasible, with respect to computational cost and available data storage facilities.
  • These settings, in conjunction with ligand pre-optimization, provide optimal performance
Summary
  • Introduction:

    Searching databases with the aim of finding structures that are similar in some fashion to a given query structure is of great interest in ligand-based design
  • Such searches are typically based on either molecular graph techniques (2D) [1] or superposition (3D) methods [2].
  • The great advantage of using 3D over 2D methods is that they explicitly take shape-dependent properties into account
  • In this way, 3D methods increase significantly the probability of finding structures which have similar shape and chemical properties to the query structure, but which are less intuitively related according to chemical class.
  • A pre-requisite for obtaining reliable results when using pre-calculated conformations (e.g. OMEGA and Catalyst) [5], is that representations of the ligands bioactive conformation must be present in the multi-conformer database
  • Objectives:

    The aim of the present study was to investigate whether modifying the parameters which control the OMEGA conformation ensemble size can generate ensembles in which there is a higher probability that the bioactive conformation is present.
  • Methods:

    OMEGA uses a depth-first searching algorithm for generating conformational ensembles.
  • It is a rule-based method that generates conformations extremely rapidly.
  • OMEGA disassembles the molecule into fragments of rotatable bonds, and reassembles the fragments based on the sorted order of the fragment energies—the depth-first method.
  • OMEGA uses a modified version of the Dreiding force field, which does not include any electrostatic terms.
  • Note that the input bond lengths, bond angles
  • Results:

    At 1.0 Å the hit-rate doubles when the input structure to OMEGA is from Corina, whereas importantly there is almost no change in the hit-rate when the pre-optimization step is carried out.
  • This graph makes it possible to compare the hit-rate to other studies, in which a different criterion to specify a match to a bioactive conformation is chosen
  • Conclusion:

    The parameters which control the nature of the conformational ensembles were varied in a statistical manner in order to optimize the performance of OMEGA for the purposes of reproducing bioactive conformations of protein-bound ligands.
  • The authors recommend setting the GP ENERGY WINDOW parameter to a low value (≤5 kcal/mol), the GP RMS CUTOFF parameter to a low value (≤0.6 Å), and generating as large conformational ensembles as feasible, with respect to computational cost and available data storage facilities.
  • These settings, in conjunction with ligand pre-optimization, provide optimal performance
Tables
  • Table1: The ligand studied, the PDB code, some crystallographic parameters and the protein in the ligand–protein complex
  • Table2: The settings, GP ENERGY WINDOW (NRG), GP NUM OUTPUT CONFS (NUM), GP RMS CUTOFF (RMS), vdW1–4, ring library (Ringlib), and the results for the 42 different runs, showing the hits, the average RMSD value (Av. RMSD) for the best-fit for each ligand, the average number of conformations (Av. Num) generated and the overall timings
  • Table3: The results, performed for each input structure, from three runs predicted to be among the very best; one “high nrg”—the best according to the statistics, and one “suggested” using a tight energy cut-off and one “small” using a tight energy cut-off as well as smaller ensemble sizes
  • Table4: The result for the database searches, showing the rank, the shape Tanimoto for the best-fit for each query ligand and the corresponding RMSD value
Download tables as Excel
Funding
  • The helpful advice from Drs Andrew Grant, Jens Sadowski, Matthew Stahl and Morten Langgård is gratefully acknowledged
Study subjects and analysis
data: 2352
None of the models with scrambled data were found to have any predictive power (Q2 < 0.2), thus ruling out the possibility of chance correlations. The PLS models were used to predict the full candidate set (n = 2352) to find and validate potentially optimal settings within and outside the selected 42 settings for each input structure. The most favorable settings were found to be similar for all four methods of generating initial conformations, with the exception of toggling the use of the ring library

Reference
  • G.M. Downs, P. Willett, Similarity searching in databases of chemical structures, in: K.B. Lipkowitz, D.B. Boyd (Eds.), Reviews in Computational Chemistry, vol. 7, VCH Publishers, New York, 1995, pp. 1–66.
    Google ScholarLocate open access versionFindings
  • (a) Y.C. Martin, 3D database searching in drug design, J. Med. Chem. 35 (1992) 2145–2154; (b) A.C. Good, J.S. Mason, Three-dimensional structure database searches, in: K.B. Lipkowitz, D.B. Boyd (Eds.), Reviews in Computational Chemistry, vol. 7, VCH Publishers, New York, 1995, pp. 67–117.
    Google ScholarLocate open access versionFindings
  • T.E. Moock, D.R. Henry, A.G. Ozkabak, M. Alamgir, Conformational searching in ISIS/3D databases, J. Chem. Inf. Comput. Sci. 34 (1994) 184–189.
    Google ScholarLocate open access versionFindings
  • T. Hurst, Flexible 3D searching: the directed tweak technique, J. Chem. Inf. Comput. Sci. 34 (1994) 190–196.
    Google ScholarLocate open access versionFindings
  • (a) P.W. Sprague, R. Hoffman, Catalyst pharmacophore models and their utility as queries for searching 3D databases, in: H. Van de Waterbeemd, B. Testa, G. Folkers (Eds.), Computer-Assisted Lead Finding and Optimization, VHCA, Basel, 1990, pp. 230–240; (b) J.A. Grant, M.A. Gallardo, B.T. Pickup, A fast method of molecular shape comparison. A simple application of a Gaussian description of molecular shape, J. Comp. Chem. 17 (1996) 1653– 1666.
    Google ScholarLocate open access versionFindings
  • J. Boström, Reproducing the conformations of protein-bound ligands: a critical evaluation of several popular conformational searching tools, J. Comput. Aided. Mol. Des. 15 (2001) 1137–1152.
    Google ScholarLocate open access versionFindings
  • OMEGA (version 1.0b4), OpenEye Science Software, 3600 Cerrillos Road, Suite 1107, Santa Fe, USA, 2001.
    Google ScholarFindings
  • ROCS (version 1.0), OpenEye Science Software, 3600 Cerrillos Road, Suite 1107, Santa Fe, USA.
    Google ScholarFindings
  • F. Bernstein, T.F. Koetzle, G.J.B. Williams, E.F. Meyer Jr., M.D. Brice, J.R. Rodgers, O. Kennard, T. Schimanouchi, M.J. Tasumi, The protein data bank: a computer-based archival file for macromolecular structures, J. Mol. Biol. 112 (1977) 535–542.
    Google ScholarLocate open access versionFindings
  • M. Hendlich, Acta Crystallogr. D54 (1998) 1178–1182.
    Google ScholarFindings
  • Jens Sadowski, AstraZeneca R&D Mölndal, Mölndal, Sweden (personal communication).
    Google ScholarFindings
  • J. Sadowski, J. Gasteiger, G. Klebe, Comparison of automatic three-dimensional model builders using 639 X-ray structures, J. Chem. Inf. Comput. Sci. 34 (1994) 1000–1008.
    Google ScholarLocate open access versionFindings
  • Modde (version 6.0), Umetrics, P.O. Box 7960, Umeå, Sweden.
    Google ScholarFindings
  • M.E. Johnson, C.J. Nachtsheim, Some guidelines for constructing exact D-optimal designs on convex design spaces, Technometrics 25 (1983) 271–277.
    Google ScholarLocate open access versionFindings
  • S. Wold, A. Ruhe, H. Wold, W.J. Dunn III, The collinearity problem in linear regression. The partial least squares (PLS) approach to generalized inverses, SIAM J. Sci. Stat. Comput. 5 (1984) 735– 743.
    Google ScholarLocate open access versionFindings
  • Simca (version 4.5), Umetrics, P.O. Box 7960, Umeå, Sweden.
    Google ScholarFindings
  • Corina Molecular Networks, GmbH Computerchemie Langemarckplatz 1, Erlangen, Germany, 2000.
    Google ScholarFindings
  • F. Mohamadi, N.G.J. Richards, W.C. Guida, R. Liskamp, M. Lipton, C. Caufield, G. Chang, T. Hendrikson, W.C. Still, MacroModel (version 7.1)—an integrated software system for modeling organic and bioorganic molecules using molecular mechanics, J. Comput. Chem. 11 (1990) 440–467.
    Google ScholarLocate open access versionFindings
  • MDDR—A Structural Database, MDL Information Systems Inc., Prous Science Publishers.
    Google ScholarFindings
  • S. Wold, Cross-validatory estimation of the number of components in factor and principal components models, Technometrics 20 (1978) 397–405.
    Google ScholarLocate open access versionFindings
  • J. Boström, P.-O. Norrby, T. Liljefors, Conformational energy penalties of protein-bound ligands, J. Comput. Aided Mol. Des. 12 (1998) 383–396.
    Google ScholarLocate open access versionFindings
  • Andrew Grant, AstraZeneca R&D Alderley Park, UK, http://www.eyesopen.com/products/shape.html (personal communication).
    Findings
Your rating :
0

 

Tags
Comments
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn
小科