| NEWS | R Documentation |
Fixed vignette compilation errors on Solaris 10.
getAttractors() now allows for plotting all attractors inside one plot. Fixed minor bugs.
Fixed compilation errors on Windows.
getAttractors() and simulateSymbolicModel() now support the identification of attractors in large networks based on a formulation as a satisfiability (SAT) problem.
plotAttractors() and plotSequence() now plot the genes in the order of the network instead of the reverse order by default. This behaviour can be controlled using a new parameter reverse which is also available in attractorsToLaTeX() and sequenceToLaTeX().
Bugfix regarding negated temporal predicates in simulateSymbolicModel().
Fixed undefined behaviour in markovSimulation().
Fixed memory misalignment in simulateSymbolicModel().
loadSBML() now accepts nodes that are constant, but have no initial value.
Support of temporal networks, and inclusion of a new simulator simulateSymbolicModel() to simulate these networks. Related functions include truthTableToSymbolic() and symbolicToTruthTable() to convert networks between the symbolic representation used by the new simulator and the truth table representation employed by the standard simulator.
New function perturbTrajectories() to assess the robustness of networks to noise in the states.
loadNetwork() can now load networks in a symbolic representation and with temporal extensions. loadSBML() and loadBioTapestry() can load symbolic networks without temporal extensions.
Most functions of the package have been adapted to work either with the symbolic representation or with the truth table representation of networks.
plotSequence() and sequenceToLaTeX() now visualize the attractor.
reconstructNetwork() now supports the specification of prior knowledge in form of required or excluded dependencies. Furthermore, it can now reconstruct networks from perturbation time series. By default, the function now returns incomplete functions with "don't care" values" instead of enumerating all possible functions.
generateTimeSeries() can now generate artificial perturbation data with simulated knock-out or overexpression of genes.
generateRandomNKNetwork() can now be supplied with a user-defined generation function for the transition functions. Generation functions generateCanalyzing() and generateNestedCanalyzing() for canalyzing functions and nested canalyzing functions are included in the package.
testNetworkProperties() supports several new tests that perturb the network states instead of the networks themselves. These are available in the test functions testAttractorRobustness() and testTransitionRobustness().
Fixed issues preventing the use of BoolNet on Big Endian systems.
Eliminated some bad style code.
Fixed some valgrind notes.
Minor bugfixes in loadNetwork().
Fixed undefined behaviour warnings for GCC 4.9.
Bugfixes in plotAttractors() and plotSequence().
Fixed compatibility issues with R 3.0 alpha.
Support for SBML: loadSBML() and toSBML() import from and export to SBML documents with the sbml-qual extension package.
saveNetwork() stores networks in the BoolNet file format.
The DNF generator employed by generateRandomNKNetwork() and simplifyNetwork() (as well as by the new functions saveNetwork() and toSBML()) now supports minimizing the canonical DNFs.
BoolNet now supports the modified interface of igraph 0.6 in all plotting functions, but is still compatible with older versions of igraph.
loadNetwork() supports comment lines in the network files.
generateTimeSeries() generates random state sequences from an existing network.
plotSequence() and sequenceToLaTeX() plot and export sequences of states similar to plotAttractors() and attractorsToLaTeX().
getAttractorSequence() extracts the states of a single synchronous attractor from an attractor information structure as a data frame.
generateState() provides a simple way to specify network states using partial assignments.
getPathToAttractor() has an additional parameter includeAttractorStates specifying which attractor states should be included in the path. The default behaviour has been changed to include all attractor states.
generateRandomNKNetwork() now supports the generation of networks using specific classes of functions. For this purpose, it has two new parameters validationFunction and failureIterations.
By default, loadNetwork() no longer changes gene names to lower case. If this behaviour is desired, it can be reactivated using the new lowercaseGenes parameter.
stateTransition() now names the state vector using the gene names.
plotAttractors() has a new parameter drawLegend to disable the legend.
The randomChainLength parameter of getAttractors() now defaults to 10000.
getAttractors(), reconstructNetwork() and markovSimulation() can now be interrupted by the user.