The dairy industry is an important part of the Australia’s agricultural industry and the CRC for Innovative Dairy Products is established to bring dairy farmers and academic researchers together, with the aim of developing methods to improve the quantity and quality of dairy yields through various methods, including, improvised breeding programs. Research efforts over the last four decades, using various model organisms, ranging from eutherians to marsupials, had given us substantial knowledge on the genetic factors regulating lactation. Most of this knowledge is recorded in published literature. Despite the presence of knowledge, a global model of genetic regulation, based on system biology, had not been established. Using mouse as the model organism, the goal of this project aims at developing a hypothetical model of murine lactation using the published literature.
Currently, MEDLINE had indexed more than 13 million citations of published works over the years, and is increasing at an increaing rate. A majority of the indexed citations contains the abstract to the paper, which can be seen as a more condensed form of knowledge represented. These citations, with abstracts, formed the primary source of knowledge where this project will draw upon. Previous efforts by a number of institutes had established curated databases of various nature from the literature, including, controlled vocabilaries (such as, Gene Ontology and Medical Subject Headings), biochemical interactions (such as Enzyme Commission), molecular evolution (such as, Cluster of Orthologous Genes) and consolidation of databases (such as, Kyoto Encyclopedia of Genes and Genomes).
Although curated databases exist but they are relatively volatile and distributed. Hence, keeping current with these databases is often time consuming and unproductive. The first step in this project will look into a means of automated assembly and updating of murine-specific subset of these databases, known as “MouseBook” in this project.
The biochemical aspect of MouseBook will be analyzed for assembly into a primitive networked model of biochemical reactions assuming homogeneity, known as “MouseWay” in this project, using stoichastic physical process modeling. At this stage, MouseWay suffers two major drawback, firstly, it takes no account of basic biological concents, such as, compartmentalization and localization. Secondly, it represents biochemical network of a universal murine call; thus, no concept of cell or tissue type. The rest of this project is targeted towards refining MouseWay, with emphasis on lactation, by eliminating the drawbacks described above via a semi-automated cyclic process of information extraction from the literature, remodeling and re-simulation.
Information extracted from the literature will be used for the following refinement of MouseWay. Firstly, a layer of cellular architecture will be superimposed on MouseWay, giving it an elementary structure of comaprtmentalization and localization. Secondly, positional information, with respect to the cellular environment, of mobile proteins, such as importins will be identified and mapped. Thirdly, knowledge about protein scaffoldings will be mined from the literature and added to the model, giving it fine structural details. At this stage, the first drawback mentioned above is largely eliminated.
The second phase of knowledge mining from literature aims to “specify” the universal murine cell into a lactogenic (mammary) cell using a reduction-addition approach, which is a cyclic process of reduction and addition. In the reduction phase, knowledge is mined specifically to identify biochemical processes that are known to the non-existent in the mammary epithelial cell. In the addition phase, biochemical processes that exist in the mammary epithelial cell but absent in the model are added.
After eliminating the drawbacks of MouseWay, the model (MouseWay) will be evolved into “MurioLactoLand: a simulatable genetic regulation model of murine lactation”, which is likely to be the eventual title for this project thesis.