Modeling Regulation Of Transcription Initiation

Author :
Publisher :
ISBN 13 : 9780542228834
Total Pages : pages
Book Rating : 4.2/5 (288 download)

DOWNLOAD NOW!

Download or read book Structured Modeling of Mammalian Transcription Networks written by and published by . This book was released on 2005 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: High-throughput gene expression measurement technologies and complete mammalian genome sequences are resources that may be instrumental in achieving the goal of systems level understanding of neuronal behavior. Initial steps towards utilizing these resources to meet this goal were taken in the present work. Systems engineering techniques were employed to address the problem of identifying from experimental data gene regulatory networks in mammalian systems. As revealed by simulation studies, the scale and complexity of biological systems make it virtually impossible to make progress on this problem through conventional application of classical system identification methods. An alternative approach, the structured modeling of transcription networks, was developed that renders the problem tractable by integrating dynamic modeling with biological constraints and the utilization of multiple data sets. The aforementioned simulation studies, the development of the structured modeling approach, and the application of the structured approach to a specific neuronal system comprise the greater part of this dissertation. Additionally, analyses of two specific biological system computational models were performed that generate hypotheses about the cellular functional relevance of the network structures of these systems. Simulation studies of gene regulatory network computational models were employed to gain insight into the challenges of gene regulatory network identification. The results of identifiability analyses, model identification studies, and network identification studies of gene network models revealed the inherent difficulty in approaching gene network identification as a straightforward system identification problem, particularly when high throughput measurements of gene expression (microarrays) are relied upon as the sole data source. It will be very difficult in general to identify gene network structure solely through modeling microarray data. This is primarily because of the high likelihood that many incorrect networks will fit the data at least as well as the correct network, rendering the determination of the correct network in the experimental setting very difficult. Furthermore, even in the ideal case where the network structure is known in advance, successful determination of model parameters may require high quality data that is difficult to obtain experimentally, involving complex perturbations and being contaminated by very little noise. A structured approach to gene network identification was developed to overcome the challenges illuminated by the simulation studies. Making two biologically motivated assumptions permitted derivation of candidate regulatory network structures by searching co-regulated gene group promoters for oven-represented transcriptional regulatory elements. These assumptions were (1) that gene regulation occurs predominantly through transcription initiation regulation, and (2) that genes with similar patterns of expression are jointly regulated ("co-expression implies co-regulation"). Utilizing the candidate network structures to define regulatory interactions, system identification techniques were used to fit dynamic models linking regulating transcription factors to target genes. Successful model fits were strongly indicative of regulatory interactions, being supported by multiple data types and being consistent with biological constraints. This structured modeling approach was successfully applied to data from the yeast cell cycle as a case study. Consideration of gene to gene variation in dynamic parameters revealed experimental conditions for which co-regulated mammalian genes will not necessarily be co-expressed, invalidating assumption (2), above. A simple technique to partially account for this variation was described. (Abstract shortened by UMI.).