Abstract:

With the increasing volume of genomic data available, it has become clear 
that biologists need computational methods for data organization and 
analysis. To develop computer applications that aid in this task, we first 
need a knowledge model that can represent biological systems. In this talk, 
I will present a knowledge-based environment for organizing biological  
data into computer-interpretable and human-browsable formats, which I 
developed during my post-doc at Stanford. This model bridges the gap between 
high-level physiological processes and molecular-level functions. My model 
enables verification of safety and soundness, which in the context of  
biological systems, may aid in predicting system behavior in the presence 
of dysfunctional processes or structural components. The model also supports 
queries that can assist in discovering relationships among processes 
and structural components that participate in them. I tested the knowledge 
model by representing the process of host cell invasion by Malaria parasites.
I assessed thirteen distinct process models that were developed in different 
fields with respect to their appropriateness for representing biological 
systems. In developing my framework, I combined the best aspects of two 
of the models: (1) Transparent Access To Multiple Biological Information 
Sources (TAMBIS) - a biological concept model, and (2) a workflow model 
that can represent the ordering of processes, the structural components 
that participate in them, and the roles that they play. The Workflow model 
maps to Petri Nets, allowing verification of properties such as boundedness 
and soundness, and determination of reachability. I composed queries 
that can aid discovering relationships among processes and structural 
components. I used reachability analysis to answer queries that relate to  
dynamic aspects of the model.  

Keywords: modeling, knowledge representation, ontology, biological process 
models, process models, Petri Nets, Workflow, Malaria

For more details, please see: 

http://smi-web.stanford.edu/projects/helix/pubs/process-model/