Protein-mRNA network Parameter Optimization

This projects fits the parameters of a regulatory network

Background Data

• mRNA levels can be measured in a variety of ways.
• The same methods can be applied to all sequences.
• Protein levels are much harder to measure
• Mutants give information
• It is possible to determine where on the DNA a transcription factor binds. Likewise to determine if two proteins interact, etc.
• Mutations are useful to: reveal components involved in a system, indicate causality, hint at how things are regulated.

System Description

• A/B/C form a loop.
• A activates transcription of B.
• B activates transcription of C.
• A represses B’s activation of C.
• C represses transcription of A.
• C increases degradation of B (which is otherwise slow).
• Other components are degraded by dx/dt = -kx.
• D is regulated by A, B and/or C but the relationship is unknown
• Hill coefficients in the system are small integers (1 or 2)

The network showed in the next figure

The equation system proposed is:

To solve it simmulated annealing and genetic algorithms were applied. An example of genetic algorithms can be found .

Data

The Experimental data that needs to be fit is shown in the figures bellow:

When mD is not considered

When mD is considered

Overexpression mutants, transcription.

Results

Using simulated annealing the results obtained are as follows: