CHE.496/2009/Responses/a13

Biological Networks

• Discussion leader: Patrick

Patrick's Response

• Biological Networks
  • A recurring point made in the paper is that biological networks are similar to other complex networks. We can extend our understanding of other complex networks such as hierarchal networks and the tools used to study those networks. Such as graph theory, in which they use to describe metabolic networks with genes as nodes and pathways as edges. Other networks the paper mentions that could be studied further in depth using these techniques is the protein-protein interaction network, transcriptional regulatory network, and metabolic networks. The paper goes through the techniques being applied to elucidate these biological networks such as ChIP for protein-DNA networks. Furthermore, the authors describe a manner of compartmentalizing biological networks by defining modules as discrete units of function separable from the whole and there are different approaches to building modules as described by the paper. In addition, motifs are defined as a set or genes or gene products with specific molecular function arranged together to perform “useful” behavior. Motifs are not modules in that they are not discrete functions separable from the whole. One could think of motifs as something similar to genetic circuits and modules as something similar to the operon structure of the bacterial gene regulatory network. Studying biological networks that interact with other networks greatly complicates analysis, but in the Network Benchmarking Article mentioned in the previous assignment. It is possible to isolate a synthetic biological network and control the inputs as to understand how the network functions. It may come to the point that in order to study very complex networks, pieces of the network may have to be built synthetically and studied and then the pieces are brought together to complete the network. Since each component of the network has been studied, understanding the entire network has been simplified, especially if you can control how intermediate products move through the network as demonstrated by Cantone.
• Programming and Engineering Biological Networks
  • The article goes through notable biological networks that have already been developed. For example: the toggle-switch in mammalian cells, pattern-forming circuits, and the synthetic riboregulator. Furthermore, an interesting topic to discuss in the discussion is the use of post-transcriptional control of gene expression utilizing synthetic RNA’s (we can include material from Dan’s paper in this from Project 1?). Christina Smolke’s antisense agents are another interesting tool for gene expression, where a targeted protein has its translation inhibited by the binding of oligo-nucleotides to the target mRNA.
• Patrick Gildea 05:06, 6 April 2009 (EDT):

Thaddeus Webb's response

Biological Networks

• Several molecular interaction networks are being elucidated including
  • protein-protein interactions
  • Transcriptional regulatory network
  • Small molecule metabolism
• High Throughput data analysis is giving picture of genetic parts.
• There is a significant gap between sequencing data and reality
• High throughput techniques
  • Two hybrid, the domains of a transcriptional activator are cut in half and fused to random proteins. When proteins interact activator works and activates reporter.
  • MS based, purify complex using affinity chromatography and identify with MS
• Results from techniques did not overlap as much as expected.
• Huge number of false positives up to 50%.
• Only considering interactions predicted by both techniques reduced false positive but decreased coverage.
• Proteome chips, fix labeled protein to chip. Pass extract over it and see what binds.
• ChIP used to find protein DNA interaction
• Biological networks share features of general complex networks. Both modular and scale free properties.
• Although biological and engineered systems appear similar in topology node function is distinct.
• A simple model of domain creation predicted domain classes created at constant rate and class members made by duplication matched E. Coli and Yeast data.
• Network topology and genome sequence data can be combined to examine the evolutionary impact of network structure on proteins.
• Proteins that interact evolve at same rate through co-evolution.
• Genes sharing regulatory motifs are co-regulatory.
• Interactions between different networks can complicate study.
• Engineering de novo is a good test of our knowledge.

Thaddeus Webb 09:46, 6 April 2009 (EDT)