Biorithm  1.1
Public Member Functions | Public Attributes
MyGainParams Struct Reference

List of all members.

Public Member Functions

 MyGainParams ()
 Constructor.
virtual ~MyGainParams ()
 Destructor.
string getAnnotationsFile () const
 Return the name of the file containing functional annotations.
bool getApplyTruePathRuleDownward () const
string getGOFile () const
 Return the name of the file containing the definition of GO.
vector< string > getFLNFiles () const
 Return the names of the file containing the functional linkage network.
string getDataIntegrationType () const
 Return the type of data integration GAIN should perform.
unsigned int getMaximumGoDepth () const
unsigned int getMinimumGoDepth () const
unsigned int getMaximumAnnotatedGenes () const
unsigned int getMinimumAnnotatedGenes () const
bool getDoNotReduce () const
 Return true if the FLN should not be reduced.
unsigned int getNumRounds () const
 Return the number of times to push flow in the FunctionalFlow algorithm.
unsigned int getNumRuns () const
bool getAnnotationsAreOriginal () const
 Return whether the annotations in the functional file on the command line are transitively closed.
string getExperimentName () const
 Return the name of the experiment provided by the --experiment-name command-line argument.
void setGroupFunctionsMethod (string method)
bool getGroupFunctionsMethodIsParent () const
bool getGroupFunctionsMethodIsDepth () const
string getGroupFunctionsMethod () const
void setFunctionEdgeProbabilities (const map< BioFunction, MyHistogram > &hists)
 Store the histograms storing mapping edge weights to probabilities for each function.
MyNT getMinimumConfidence () const
 Return the minimum prediction confidence to consider.
MyNT getOneVersusNoneSinkSourceArtificialEdgeWeight () const
bool processCategory (string cat) const
 Return true if and only if cat is a function category that GAIN should process.
string getOutputDirectory () const
 Return the name of the directory to print results to.
string getOutputFileName (string name, string dir="") const
const set< string > * getEvidenceCodesToIgnore () const
 Return pointer to set of evidence codes to ignore.
void printInvocation ()
bool getPrintDetailedCVResults () const
 Return true if detailed cross validation results should be printed.
bool getRunGraphviz () const
 Return true if the user wants to run the Graphviz algorithms.
bool getPerformSanityCheck () const
string getVisualiseParamsFile () const
int getNumPredictionsToPrintPerFunction () const
 Return the number of predictions to be printed for each function.
bool runAnyOneVersusNoneAlgorithm () const
 Return true if user asked for any one-versus-none algorithm to be executed.
bool runOneVersusAllAlgorithm (string algorithm) const
bool runOneVersusNoneAlgorithm (string algorithm) const
string getPredictionsFile () const
string getValidationAnnotationsFile () const
 Return the name of the file containing (new) annotations to be used to validate predictions.
string getLibSVMDirectory () const
 Return the name of the directory containing the libSVM executables.
string getSVMLightDirectory () const
 Return the name of the directory containing the SVMLight executables.
string getLibSVMTrainOptions () const
 Return a string containing extra options to be passed to the libSVM training programme.
string getLibSVMTestOptions () const
 Return a string containing extra options to be passed to the libSVM testing programme.
string getSVMLightTrainOptions () const
 Return the name of extra options to be passed to the SVMLight training programme.
string getSVMLightTestOptions () const
 Return the name of extra options to be passed to the SVMLight testing programme.
bool getComputeTreewidth () const
 Return true if GAIN should compute the treewidth of the FLN.
bool getWeightEvidenceCodes () const
 Return true if evidence codes should be weighted.
string getEvidenceCodeWeightsFile () const
 Return the name of the file that contains evidence code weights.
void setParameters (gengetopt_args_info &gainOptions)

Public Attributes

ofstream cvStream
ofstream ecwCVStream
ofstream detailedCVStream
ofstream edgeWeightsStream
ofstream geneUniverseStream
ofstream groupedCVStream
ofstream groupedEcwCVStream
ofstream invocationStream
ofstream logStream
ofstream predictionsStream
ofstream propagationDiagramsStream
ofstream sanityCheckStream
ofstream statsStream
map< string, ofstream * > outputStreams
string _invocationFile
bool allowZeroStates
MyNT gainThreshold
bool gainThresholdUserInput
bool gainThresholdRangeUserInput
string geneExpressionFile
set< string > ignoredThings
set< string > ignoredEvidenceCodes
bool justUseCorrelations
MyGainUpdateType updateType
bool useDegree
unsigned int localRuleDistance
bool computePvalues
bool checkPropagation
bool crossValidate
string crossValidationFile
bool leaveOneOutCrossValidate
int foldCrossValidate
MyGainFoldInfo foldInfo
MyGainLOOCVInfo loocvInfo
MyNT functionFrequency
bool printAllStates
bool unclampPositives
bool unclampNegatives
bool useStateWeights
bool visualisePredictions
bool visualiseCrossValidation
bool visualiseCut
bool onlyPredictions
bool onlyCrossVal
bool weightEdgeTypesCutoff
string cutoffsFile
bool weightEdgeTypesDepth
bool weightEdgeTypesJaccard
bool groupEdgeTypes
string edgeTypeGroupsFile
bool useCustomRNGSeed
int customRNGSeed
string edgeWeightingScheme

Member Function Documentation

Return true iff the GO true path rule should be applied downward to the annotations.

bool MyGainParams::getDoNotReduce ( ) const [inline]

Return true if the FLN should not be reduced.

"Reduction" of an FLN is the operation of deleting all connected components that do not contain both positive and negative examples. This operation does not make sense for one-versus-none algorithms, which do not have any negative examples. Hence, this method returns false if even one algorithm is invoked on the command line with the --one-versus-none option.

unsigned int MyGainParams::getMaximumAnnotatedGenes ( ) const [inline]

Return the maximum number of annotated genes for which GAIN should analyse (make predictions or perform cross validation for) a GO function.

unsigned int MyGainParams::getMaximumGoDepth ( ) const [inline]

Return the maximum depth at which GAIN should analyse (make predictions or perform cross validation for) a GO function.

unsigned int MyGainParams::getMinimumAnnotatedGenes ( ) const [inline]

Return the minimum number of annotated genes for which GAIN should analyse (make predictions or perform cross validation for) a GO function.

unsigned int MyGainParams::getMinimumGoDepth ( ) const [inline]

Return the minimum depth at which GAIN should analyse (make predictions or perform cross validation for) a GO function.

unsigned int MyGainParams::getNumRuns ( ) const [inline]

Return the number of times to run the algorithm (semi-hierarchical Hopfield or hierarchical Hopfield).

Return the value passed as the command-line argument ovn-sinksource-edge-weight.

bool MyGainParams::getPerformSanityCheck ( ) const [inline]

Return true if the user wants to perform a sanity check, e.g., whether gene ids in gene expression data match the ones in the functional annotations file.

string MyGainParams::getPredictionsFile ( ) const [inline]

Return the name of the file containing predictions made by GAIN during an earlier run or by another algorithm.

string MyGainParams::getVisualiseParamsFile ( ) const [inline]

Return the name of the file specifying how nodes and edges should look in propagation diagrams.

Print details about the invocation in a two-column format.

Each row contains the name of a parameter and its value.

bool MyGainParams::processCategory ( string  cat) const [inline]

Return true if and only if cat is a function category that GAIN should process.

Parameters:
catA string representing a biological function category, e.g., "biological process" or "p".

If the user did not specify any function categories in the command line, the method always returns true.

bool MyGainParams::runOneVersusAllAlgorithm ( string  algorithm) const

Return true if user asked for algorithm to be executed.

Parameters:
[in]algorithmThe name of a one-versus-all algorithm, e.g., "Hopfield" or "sinksource".
bool MyGainParams::runOneVersusNoneAlgorithm ( string  algorithm) const

Return true if user asked for algorithm to be executed.

Parameters:
[in]algorithmThe name of a one-versus-none algorithm, e.g., "functional-flow"

Set all parameters based on the command line.

Parameters:
gainOptionsan instance of gengetopt_args_info, created by a call to cmdline_parser.

The documentation for this struct was generated from the following files:
 All Classes Functions Variables Typedefs Friends