Brian D. Hudson

Strategic Pooling of Compounds for High-Throughput Screening

Bringing new medicines to the market depends on the rapid discovery of new and effective drugs, often initiated through the biological testing of many thousands of compounds in high-throughput screening (HTS). Mixing compounds together into pools for screening is one way to accelerate this process and reduce costs. This paper contains both theoretical and experimental data which suggest that careful selection of compounds to be pooled together is necessary in order to reduce the risk of reactivity between compounds within the pools.

Toward high throughput 3D virtual screening using spherical harmonic surface representations

Searching chemical databases for possible drug leads is often one of the main activities conducted during the early stages of a drug development project. This article shows that spherical harmonic molecular shape representations provide a powerful way to search and cluster small-molecule databases rapidly and accurately. Our clustering results show that chemically meaningful clusters may be obtained using only low order spherical harmonic expansions. Our database search results show that using low order spherical harmonic shape-based correlation techniques could provide a practical and efficient way to search very large 3D molecular databases, hence leading to a useful new approach for high throughput 3D virtual screening. The approach described is currently being extended to allow the rapid search and comparison of arbitrary combinations of molecular surface properties.

Pattern recognition display methods for the analysis of computed molecular properties

SummaryPattern recognition methods, particularly the ‘unsupervised learning’ techniques, are well suited for the preliminary analysis of the large data sets produced by computer chemistry. The use of linear and non-linear display methods for such exploratory analysis are exemplified with the aid of two data sets of biologically active molecules. Advantages and disadvantages of these techniques are discussed.

Biological data mining with neural networks: implementation and application of a flexible decision tree extraction algorithm to genomic problem domains

In the past, neural networks have been viewed as classification and regression systems whose internal representations were extremely difficult to interpret. It is now becoming apparent that algorithms can be designed which extract understandable representations from trained neural networks, enabling them to be used for data mining, i.e. the discovery and explanation of previously unknown relationships present in data. This paper reviews existing algorithms for extracting comprehensible representations from neural networks and describes research to generalize and extend the capabilities of one of these algorithms. The algorithm has been generalized for application to bioinformatics datasets, including the prediction of splice site junctions in Human DNA sequences. Results generated on this datasets are compared with those generated by a conventional data mining technique (C5) and conclusions drawn.

A computer-aided drug discovery system for chemistry teaching.

The Schools Malaria Project (http://emalaria.soton.ac.uk/) brings together school students with university researchers in the hunt for a new antimalaria drug. The design challenge being offered to students is to use a distributed drug search and selection system to design potential antimalaria drugs. The system is accessed via a Web interface. This e-science project displays the results of the trials in an accessible manner, giving students an opportunity for discussion and debate both with peers and with the university contacts. The project has been implemented by using distributed computing techniques, spreading computer load over a network of machines that cross institutional boundaries, forming a grid. This provides access to greater computing power and allows a much more complex and detailed formulation of the drug design problem to be tackled for research, teaching, and learning.

Conformation-Dependent QSPR Models: logPOW

Quantitative structure-property relationships for predicting the water-octanol partition coefficient, logP(OW), are reported. The models are based on local properties calculated at the standard isodensity surface using semiempirical molecular orbital theory and use descriptors obtained as the areas of the surface found in each bin in a predefined binning scheme. The effect of conformation is taken into account but was found to have little effect on the predictive power of the models. A detailed error analysis suggests that the accuracy of the models is limited by that of the experimental data and that the best possible performance is approximately ±0.5 log units. The models yield a local hydrophobicity function at the surface of the molecules.

QSAR studies of the pyrethroid insecticides. Part 3. A putative pharmacophore derived using methodology based on molecular dynamics and hierarchical cluster analysis.

Previous studies of the conformational behaviour of a group of synthetic pyrethroid insecticides have been extended to a more structurally diverse set. This includes compounds with different backbones and differing stereochemistry, with both Types I and II biological activity. These compounds also encompass a large range of biological activities. A parameterisation of the CHARMM force field for these compounds has been performed and the extra parameters are reported. Conformational sampling, using molecular dynamics (MD), has been performed for each of the 41 active structures. The accessible conformations of each have been characterised by the values of the common torsion angles using hierarchichal cluster analysis (HCA). A further CA, based on the centroids derived from the conformational sampling, identified a conformation common to at least 39 of the 41 structures. The critical torsion angles of this conformation lie at the centre of the molecule about the ester linkage and are defining an extended conformation, which differs from the minimum energy conformation of deltamethrin used previously. This may represent a putative pharmacophore for kill. The methods used here improve significantly on those used previously. The CHARMM force field was parameterised for the compounds and an improved method of conformational sampling, based on centroid clustering, has also been used.

Vicinity analysis: a methodology for the identification of similar protein active sites

Vicinity analysis (VA) is a new methodology developed to identify similarities between protein binding sites based on their three-dimensional structure and the chemical similarity of matching residues. The major objective is to enable searching of the Protein Data Bank (PDB) for similar sub-pockets, especially in proteins from different structural and biochemical series. Inspection of the ligands bound in these pockets should allow ligand functionality to be identified, thus suggesting novel monomers for use in library synthesis. VA has been developed initially using the ATP binding site in kinases, an important class of protein targets involved in cell signalling and growth regulation. This paper defines the VA procedure and describes matches to the phosphate binding sub-pocket of cyclin-dependent protein kinase 2 that were found by searching a small test database that has also been used to parameterise the methodology.

QSAR studies using the parashift system

A novel way of describing molecules in terms of their surfaces and local properties at the surfaces is described. The use of these surfaces and properties to explain chemical reactivity and model simple molecular properties has already been demonstrated. This study reports an examination of the use of these descriptions of molecules to model a simple chemical interaction (complex formation) and a diverse set of mutagens. Both of these systems have been modelled successfully and the results are discussed. †This paper is dedicated to the memory of our colleague and dear friend Martyn Ford who passed away in June 2007.

The Extraction of Information and Knowledge from Trained Neural Networks

In the past, neural networks were viewed as classification and regression systems whose internal representations were incomprehensible. It is now becoming apparent that algorithms can be designed that extract comprehensible representations from trained neural networks, enabling them to be used for data mining and knowledge discovery, that is, the discovery and explanation of previously unknown relationships present in data. This chapter reviews existing algorithms for extracting comprehensible representations from neural networks and outlines research to generalize and extend the capabilities of one of these algorithms, TREPAN. This algorithm has been generalized for application to bioinformatics data sets, including the prediction of splice junctions in human DNA sequences, and cheminformatics. The results generated on these data sets are compared with those generated by a conventional data mining technique (C5) and appropriate conclusions are drawn.