perturbTrajectories function

Perturb the state trajectories and calculate robustness measures

Perturbs the state trajectories of a network and assesses the robustness by comparing the successor states or the attractors of a set of initial states and a set of perturbed copies of these initial states.

perturbTrajectories(network, 
                    measure = c("hamming", "sensitivity", "attractor"), 
                    numSamples = 1000, 
                    flipBits = 1, 
                    updateType = c("synchronous", "asynchronous", "probabilistic"), 
                    gene, 
                    ...)

Arguments

• network: A network structure of class BooleanNetwork, SymbolicBooleanNetwork or ProbabilisticBooleanNetwork whose robustness is measured.
• measure: Defines the way the robustness is measured (see Details).
• numSamples: The number of randomly generated pairs of initial states and perturbed copies. Defaults to 1000.
• flipBits: The number of bits that are flipped to generate a perturbed copy of an initial state. Defaults to 1.
• updateType: If measure="hamming", the type of update that is performed to calculate successor states.
• gene: If measure="sensitivity", the name or index of the gene for whose transition function the average sensitivity is calculated.
• ...: Further parameters to stateTransition and getAttractors.

Details

The function generates a set of numSamples initial states and then applies flipBits random bit flips to each initial state to generate a perturbed copy of each initial state. For each pair of initial state and perturbed state, a robustness statistic is calculated depending measure:

If measure="hamming", the normalized Hamming distances between the successor states of each initial state and the corresponding perturbed state are calculated.

If measure="sensitivity", the average sensitivity of a specific transition function (specified in the gene parameter) is approximated: The statistic is a logical vector that is TRUE if gene differs in the successor states of each initial state and the corresponding perturbed state.

If measure="attractor", the attractors of all initial states and all perturbed states are identified. The statistic is a logical vector specifying whether the attractors are identical in each pair of initial state and perturbed initial state.

Returns

A list with the following items: - stat: A vector of size numSamples containing the robustness statistic for each pair of initial state and perturbed copy.

• value: The summarized statistic (i.e. the mean value) over all state pairs.

References

I. Shmulevich and S. A. Kauffman (2004), Activities and Sensitivities in Boolean Network Models. Physical Review Letters 93(4):048701.

See Also

testNetworkProperties, perturbNetwork

Examples

## Not run:

data(cellcycle)

# calculate average normalized Hamming distance of successor states
hamming <- perturbTrajectories(cellcycle, measure="hamming", numSamples=100)
print(hamming$value)

# calculate average sensitivity of transition function for gene "Cdh1"
sensitivity <- perturbTrajectories(cellcycle, measure="sensitivity", numSamples=100, gene="Cdh1")
print(sensitivity$value)

# calculate percentage of equal attractors for state pairs
attrEqual <- perturbTrajectories(cellcycle, measure="attractor", numSamples=100)
print(attrEqual$value)
## End(Not run)