| ... Presentations
|
| Digital Chemistry
staff regularly attend international conferences
and other meetings, often giving presentations about
Digital Chemistry research and development. Some
recent presentations are listed below, with links
to the publications where applicable. Copies of
the presentations are available by contacting Digital
Chemistry. |
| |
| ...8th International Conference on Chemical Structures, Noordwijkerhout, Netherlands, 1-5 June 2008 |
Opportunities for Integrating Markush Patent Searching
with Drug Discovery
John M. Barnard and Matthew Wright
|
| Information about existing chemical compounds is regularly used in designing chemical libraries for new drug discovery. The information is normally derived from a diverse range of specific structure databases, including internal, bought-in, and public ones, which are managed using Oracle RDBMS data cartridges. Integrated systems incorporate similarity searching, physicochemical property profiling, structure-based cluster analysis etc. across a variety of databases with a common user interface. There is clear potential for utilizing information about the chemical structures in patents in the drug discovery process. This could allow the avoidance of areas of chemistry already covered by patents, identification of un-patented areas of chemical space, and also analysis of the different classes of chemical structure being patented in particular therapeutic areas.
Though some limited use is made of databases of specific structures from patent claims, searching of the Markush structures currently remains the province of dedicated on-line systems with proprietary databases. Some Oracle cartridges have recently been extended to handle Markush representations of un-enumerated combinatorial libraries alongside discrete molecules. These provide a possible basis for extension to handle the more complex Markush structures found in patents. The issues involved in this extension, both technical and commercial, are discussed. |
| |
Rapid Property Profiling and Similarity Calculations
in Large Virtual Libraries
John M. Barnard and Geoff M. Downs
|
| Extremely large virtual compound libraries, containing up to 1012 or even 1015 different molecules, may be used in drug discovery, and challenge even the fastest virtual high-throughput screening (VHTS) analyses. Even using Markush structure-based property-calculation techniques, production of property profiles for simple drug-like characteristics such as the Lipinksi properties can be prohibitively slow. Appropriate sampling of library members can permit much faster analyses, and factors affecting the accuracy of property distribution profiles based on such sampling are discussed. Direct analysis of the Markush structure can also be used to calculate upper and lower bounds on the range of values for a particular property, without the need to enumerate individual values.
Exhaustive comparisons of individual library members against a target molecule, to identify the most similar, are also too time-consuming to be used on very large libraries. An approximate similarity search algorithm has been developed, which allows selection of molecules from a library that are highly similar to a specified target, though not guaranteed to be the most similar. The sets of molecules selected by this algorithm are compared to those identified by exhaustive similarity search. |
| |
| ...
IPI-Confex, Seville, Spain, 2-5
March 2008 |
Towards In-House Searching of Markush Structures from Patents
John M. Barnard |
| Well-integrated in-house ("behind the firewall") cheminformatics systems able to handle databases of individual molecules for substructure and similarity searching have been available for many years, normally in the context of Oracle RDBMS data cartridges. These systems are able to search diverse databases (including internal, bought-in, and public ones) with a single interface, and may include additional features such as physicochemical property profiling and structure-based cluster analysis. Recently some have been extended to handle Markush representations of un-enumerated combinatorial libraries alongside discrete molecules. This paper describes ongoing work to extend them further to handle the more complex Markush structures found in patents, and discusses the issues involved in doing so, and the opportunities that may arise. |
| |
| ...
ACS 232nd National Meeting, San Francisco, USA, 10-14
September 2006 |
Searching of Markush Structures and its Uses
John M. Barnard, Anthony P. F. Cook, Geoff M. Downs, Annette von Scholley-Pfab, Daniel G. Thomas, P. Matthew Wright, Jimmy Chung, Gavin Harper, and Stephen Pickett |
| Direct substructure searching of Markush structure representations of large virtual combinatorial libraries offers substantial speed advantages over searching the enumerated molecules, since each individual building block needs only to be examined once. Markush substructure searching has several potential applications in the management of collections of large virtual libraries, and has been implemented in the context of an Oracle RDBMS data cartridge. The libraries can be stored as Markush structures (or as pools of precursor molecules with generic reactions) rather than as enumerated products. The presence or absence (in the product set) of substructures or complete molecules of interest can be rapidly identified, and just the relevant molecules enumerated if required. Physicochemical and ADME properties based on additive contributions from predefined sets of substructures can be rapidly calculated. Work on a recent collaborative project will also be described, in which Markush substructure search techniques are used for rapid generation of the "reduced feature graphs" (Harper et al., J. Chem. Inf. Comput. Sci., 2004, 44, 2145-2156) present in a large virtual library, and for identification of those that contain specified patterns of feature nodes, along with the corresponding individual molecules. |
| |
... 3rd Joint Sheffield Conference on Chemoinformatics,
21-23 April 2004
Substructure
|
Better Clusters Faster
John Barnard
|
|
Abstract:
In this study, we investigate a new and faster
hierarchical clustering method, Divisive K-means,
and a new method, Stepped Partitioning, for selecting
optimal sets of clusters from a hierarchy. The efficiency
of these methods, in terms of processing time and
memory requirements, will be discussed along with
quantitative and qualitative assessments of their
effectiveness in grouping compounds into appropriate
clusters. Ways in which these two methods can be
used alongside existing methods to improve the clustering
of very large datasets will be examined. We will
also discuss the use of parallel hardware to cluster
extremely large datasets. |
| |
| ...
AccelrysWorld 2003, San Diego, CA, USA 24-26 February
2003 |
Ultra-Fast Handling
of Combinatorial Libraries
John
Barnard |
| A discussion of
BCI's software for combinatorial library processing,
including structure fingerprints generation, ADME
property and topological index calculation, structure
and substructure searching etc. Its implementation
with a flexible toolkit API allows it to be incorporated
into a variety of different user components and
applications, and the ultra-fast calculation of
descriptors can be used "on-the-fly" during
optimisation of combinatorial sub-libraries of extremely
large virtual libraries. |
| |
| ...
222nd National Meeting of the American Chemical
Society, Chicago, August 26-30, 2001 |
Identifying and Finding
Compounds in Combinatorial Libraries
John
Barnard and Geoff Downs |
| A description of
recent work using Markush structure-handling techniques
for rapid enumeration of combinatorial libraries,
and to identify the overlap between libraries. |
| |
Clustering of Very
Large Datasets
Geoff
Downs and John Barnard |
| A review of work
on cluster analysis of large chemical datasets,
and discussion of some recently-published algorithms
which might be applicable. |
| |
| ...
2nd Joint Sheffield Conference on Chemoinformatics,
April 9-11, 2001 |
Fast Descriptor Calculation
for Combinatorial Libraries
Geoff
Downs and John Barnard |
| A description of
BCI's work on rapid generation of Lipinsky properties
and topological indices for the members of combinatorial
libraries, using Markush structure-handling techniques.
|
| |
| ...
5th International Conference on Chemical Structures,
Noordwijkerhout, The Netherlands, 6-10 June 1999 |
Computer Manipulation
of Large Virtual Combinatorial Libraries for Diversity
Analysis and Subset Selection
John
Barnard and Geoff Downs |
| A description of
BCI's work on rapid generation of structure fingerprints
for the members of combinatorial libraries, using
Markush structure-handling techniques. |
|
