Skip to main content

Engineering Genes Based Neural Control in Systems Biology

Supervisors: Dr Kang Li

Only in the last few years, the field of systems biology or molecular systems biology has taken shape with the goal of unravelling the basic dynamic processes, feedback control loops, and signal processing and regulation mechanisms at the cellular level of the living creatures. This is due to a number of factors:

  1. Nowadays biologists worldwide have recognized that systems-level knowledge is urgently required in order to conceptualize and organize the revolutionary developments taking place in the biological sciences.
  2. Bioinformatics has been tremendously successful in facilitating the sequencing of the genomes and protein structure prediction in the recent years. Mathematical tools, algorithms and user-friendly software are now indispensable for biologists and pharmaceutical researchers. Many biologists have come to accept and value the use of mathematical tools to achieve the system level understanding of the biological systems.
  3. High-throughput data collection techniques such as microarray make possible the simultaneous monitoring of the activity and temporal relations of numerous genes and the concentrations of proteins and metabolites, thus allowing for the study of microscopic dynamic interactions among cellular components, and making a systems-level view of cells particularly natural. In the meantime, the huge amounts of data being generated by genomics and proteomics require new theoretical approaches to interpretation and organization.
  4. Medical advances also drive this new emphasis. Many in the pharmaceutical industry have come to realise that only by understanding cells as a whole can one identify novel targets and delivery for new drugs, and understand their systemic effects. Gene therapies will depend on a more global understanding of dynamic interactions among genes and their cellular environment.
  5. Finally, at a more philosophical level, there also is the fact that current experimental methods permit making falsifiable predictions, bringing modern biology closer to physics and chemistry as a science.

Many biological systems exist strong nonlinearity. To regulate such complex nonlinear systems, a systematic method is required to incorporate the nonlinear features uniquely identifiable from the system with improved control performance and design transparency. Differing from all existing methods for the modelling and control design of engineering systems, this project intends to establish a new method and associated algorithms for the real-time control of engineering systems. The proposed method performs system modelling and control by first establishing the 'eng-genes' - fundamental engineering functions from 'a priori' engineering knowledge. These 'eng-genes' are then composed and combined to produce a simple neural structure for modelling and control. The 'eng-genes' reflect the nonlinearity uniquely existing within the specific engineering system, therefore it is natural to compose and code these 'eng-genes' and the associated neural structure into appropriate chromosome representations which then evolve to produce a neural controller for adaptive regulation and control of the system under different operation situations.

The proposed methodology will be applied to a wide range of biological systems.