Computational Biology - Current and Recent Projects & Research Grants
- Project IQ - Inductive Queries for Mining Patterns and Models
- At present, there is a lack of a generally accepted framework for data mining - the quest for such a framework is a major research priority. The most promising approach to this task is taken by inductive databases (IDBs) that contain not only data but also local and global patterns. Project IQ aims to develop the theory and practical approaches to inductive querying of global models, as well as the answering of complex inductive queries that involve both local patterns and global models. Based on this, showcase IDBs in the area of bioinformatics will be developed that will enable users to query data about drug activity, gene expression, gene function and protein sequences. More information.
Contact: Ross King <firstname.lastname@example.org> or Amanda Schierz <email@example.com>
- Robot Scientist
- Using Closed Loop Learning to automate a laboratory robot. The robot will be used for experiments in functional genomics, and machine learning will be used to guide the experiments.
See the Robot Scientist homepage for more details.
- Novel Machine Learning Approaches to Molecular Coherent Control
- The objective is to develop the first UK capability in chemical quantum control and, in the process, develop generic active learning software for ultra-fast closed loop control applications. This project is in collaboration with Ben Whitaker from the University of Leeds.
- Bio-logical Database
- A consortium will be formed to develop intelligent database technology for functional genomics using data from S. cerevisiae (bakers yeast) as a model. Existing bioinformatic databases are generally designed only to store data and to make these data available to other biologists. We propose a reasoned but radically different type of database, one of which will have the capacity to aid biologists infer new biological knowledge. Such intelligent databases are needed to transform the flood of functional genomics data into knowledge. The database will store a curated collection of functional genomics data and background knowledge on S. cerevisiae. This database will be used to develop a logical inference tools based on: deduction, abduction, and induction (machine learning). These tools which will aid users to identify interesting new patterns and hypotheses in the data.
- HiMet - Hierarchical Plant Metabolomics for Gene Function and Mode of Action Discovery
- Exploitation of genome sequences on public databases requires integrated approaches to determine gene function. Rapid phenotype 'fingerprinting' is essential to exploit reverse genetics resources, such as the ATIS 'gene tag' Arabidopsis mutant collection. We propose a hierarchical metabolomics approach opening with machine learning computation of high-throughput analysis of plant metabolites to facilitate comprehensive gene function determination. A major output will be a metabolome 'fingerprint' database that will provide the basis of future applied post-genomic technologies for high-throughput mode-of-action analysis for agrochemical discovery as well as plant breeding, 'substantial equivalence' testing of GMOs and food quality assessment.
- ArMet - a framework for the description of plant metabolomics experiments and their results
- This facilitates the development of systems for the storage and dissemination of plant metabolomics experimental data and provides a starting point for the development of community data standards for the metabolomics community, much as MIAME has done for the microarray community and Pedro and HUPO PSI are doing for the proteomics community.
See www.armet.org for details.
- MeT-RO (Metabolomics at Rothamsted)
- An initiative to establish the UK Centre for Plant and Microbial Metabolomic Analysis. This is the result of a major BBSRC initiative to establish a critical mass of resources that can be applied to plant and microbial metabolomics. The project is co-ordinated by the National Centre for Plant and Microbial Metabolomics at Rothamsted Research, and includes important partnerships in bioinformatics and advanced computational techniques with the University of Wales, Aberystwyth and in FT-ICR-MS application to plant samples with UMIST.
See www.metabolomics.bbsrc.ac.uk for more details.
- Grid-enabled Lab Robots and the Robot Scientist
- This project will design and produce software and protocols to allow laboratory robots to be connected as part of the Grid. In biology, lab robotics is beginning to allow mundane tasks to be automated and enormous amounts of new data to be produced. In the Grid community, standards, protocols, middleware and applications are now under development. Grid-enabling the lab equipment allows the Robot Scientist software to completely control the whole loop, from hypothesis to experiment to hypothesis again.
- Gene function discovery using Genetic Programming
- Applying novel genetic programming techniques to the analysis of genome-scale datasets. Research aspects include investigation of the evolutionary representation of multi-label classification problems and combining backpropagation with GP.
- Metabolome technology for the profiling of GM and conventionally bred plant materials.
- As part of a continuing collaboration between biologists and computer scientists, this project was largely responsible for the design of ArMet (q.v.), a framework for the representation of metabolomics data. This was implemented as a web-accessible relational database. The project focused on
Arabidopsis thalianaand Solanum tuberosum(potato). The former were grown under a series of carefully controlled conditions and the latter at a field scale. GC-MS and ESI-MS analytical technologies were used on the samples to generate metabolic profiles and fingerprints. State of the art statistical and machine learning methods were applied to these data and their efficacy in safety assessment was evaluated.
- Genome Scale Prediction of Protein Function from Sequence using Data Mining
- Using ILP and decision trees to predict the function of proteins in organisms such as E. coli, M. tuberculosis, yeast (S. cerevisiae) and the plant Arabidospis thaliana. Genomes of these organisms have been sequenced and annotated, but often over a third of the functions of the genes are still unknown. Predictions for these, made by data mining, can then make experimental determination simpler. Further details for E. coli and M. tuberculosis, or yeast.
- Bioinformatic System Identification using Qualitative Models
- The aim of this project is to develop and test a tool for discovering biological knowledge in transcriptome, proteome or metabolome data, or a combination of these three. The tool will be based on identifying qualitative models of biological systems. Further information.
- Drug design
- Using Inductive Logic Programming for 3-dimensional structure based drug design. Further details.
- Protein Secondary Structure Prediction
- Using cascaded multiple classifiers for prediction in Prof. Contact: <firstname.lastname@example.org>
BBSRC: International Scientific Interchange Scheme - "Data Models for Metabolomics Data: Symbioses and Analyte/Metabolite Association" Two weeks travel and subsistence to New Zealand (AgResearch Grasslands) in February 2005 (Hardy, N.) Jan 2005, £2150.
BBSRC: travel grant to collaborate with Japan (Osaka, Kyushu). (King, R. D.) 1/2005 - 12/2009. £40,000.
Royal Commission for the Great Exhibition of 1851: "Grid-enabled Lab Robots and the Robot Scientist", (Clare, A.) 10/2004 - 9/2006, Research Fellowship.
EPSRC: "Novel Machine Learning Approaches to Molecular Coherent Control", (R. D. King, J. J. Rowland, B. Whitaker), 2/2004 - 1/2007
BBSRC: "Hierarchical plant metabolomics for gene function and mode of action discovery", (Scott I. M., Draper J., Hardy N., Goodacre R., King R., Kell D. B., Rowland J. J., Darby R. M.), £462,596. 02/2003 - 02/2006
BBSRC: "Bio-Logical: an intelligent database for knowledge discovery in functional genomics", (R. D. King, S. G. Oliver, A. Srinivasan), £591,244. 11/2002 - 11/2006
FSA: "G02006: Metabolome technology for the profiling of GM and conventionally bred plant materials." 9/2001 - 8/2004.
BBSRC: "Prediction of protein function in plant genomes using data mining", (R. D. King, H. J. Ougham, S. T. May), £164,992. 6/2001 - 4/2004
BBSRC: "Automated discovery of gene function via genomic computing of proteome data", (J. J. Rowland) Studentship. 10/2001 - 9/2004
Royal Society / Wolfson: "Laboratory Refurbishment in Bioinformatics and Functional Genomics", (R.D. King, J.J. Rowland, and G.M. Coghill) £70,965. 2/2000 - 8/2000
EPSRC: "Solving peak shift problems in infrared spectroscopy with the peak parameter representation (PPR)". (B.K. Alsberg) £62,439. 7/2000 - 6/2003
BBSRC: "Characterisation of Intact Microorganisms using Electrospray Ionisation Mass Spectrometry", (R. Goodacre, D. B. Kell and J. J. Rowland) £205,640: 8/99 - 7/02.
BBSRC: "Inductive Logic Programming for 3-Dimensional Structure Based Drug Design", R. D. King and B. K. Alsberg, £146,664, 5/1999 - 4/2002.
BBSRC: "The Robot Scientist: Application to Functional Genomics", R.D. King, J. J. Rowland, and D. B. Kell (IBS), £313,380, 4/1999-3/2002.
BBSRC: "Making the most of a genome sequence: the application of transcriptome and proteome analysis to Streptomyces...", D. B. Kell, (IBS) J.J. Rowland and B. K. Alsberg, £232,980, 3/1999 - 2/2002.
BBSRC: "Bioinformatic System Identification", R.D. King, G.M. Coghill, D.B. Kell (IBS), £160,696, 11/1998 - 10/2001 (joint with Model-based Systems Group)
BBSRC: "Functional Genomics via the Metabolome", D. B. Kell (IBS) and J.J. Rowland, £321,360, 1/1999 - 12/2002.
BBSRC: ``Improved Protein Secondary Structure Prediction Using Advanced Statistics and Machine Learning'', R. D. King, £135,946, 1998 - 2001
EPSRC: ``Deep Database Mining'', R. D. King, £53,117, 10/1997 - 9/2000.
UWA Research Fund: ``Robotic recovery of uncultured microbes'' D. B. Kell (IBS), G.W. Griffith (IBS), and J.J. Rowland £10,000, 10/1997 - 9/1998.
BBSRC (ROPA): ``The Development of Histometrics'' D. B. Kell (IBS) and J.J. Rowland, £90,544, 12/1997 - 11/1999.
EPSRC: ``Explanatory Analysis of Complex Vibrational Spectra using Genetic Programming of Fuzzy Rules'' D. B. Kell (IBS) and J.J. Rowland £94,159, 9/1997 - 8/1999.
HEFCW: ``Intelligent Systems in Complex Biological and Biotechnological Analysis - an Industrial Support Facility'' D. B. Kell (IBS) and J.J. Rowland £96,000, 1997-1999
BBSRC: ``Rapid Analyis of Multiple Determinands using Ultrasensitive, Dispersive, Raman Spectroscopy and Supervised Learning'' D. B. Kell (IBS) and J.J. Rowland, £474,536, 1996-1998
Glaxo R & D Ltd: ``An FT-IR Based Metabolic Microscope for Biotechnology'' D. B. Kell (IBS) and J.J. Rowland, £122,145, 1996-1999
BBSRC: ``An FT-IR Based Metabolic Microscope for Biotechnology'' D. B. Kell (IBS) and J.J. Rowland, £284,720, 1995-1998.
BBSRC: ``Quantification of Microbial Productivity via Multi-angle Light Scattering and Supervised Learning'' D. B. Kell (IBS) and J.J. Rowland, £159,772, 1995-1997