evolve: Running the ChIPanalyser implementation of a Genetic...

In patrickCNMartin/ChIPanalyser: ChIPanalyser: Predicting Transcription Factor Binding Sites

Running the ChIPanalyser implementation of a Genetic algorithm.

Description

evolve pushes a starting population to evolve in a genetic algorithm.

Usage

evolve(population,DNASequenceSet,ChIPScore,
       genomicProfiles,parameters=NULL,generations=100,mutationProbability=0.3,
       offsprings=5,chromatinState=NULL,
       method="geometric", lambda=TRUE,
       checkpoint=TRUE,
       filename=NULL, cores=1)

Arguments

population	numeric value describing the number of individuals in the starting population. Alternatively - a starting population list as returned by generateStartingPopulation. NOTE: if numeric - the parameter argument is also required.
DNASequenceSet	DNAStringSet object containing DNA sequences of interest (Extracted from BSgenome)
ChIPScore	ChIPScore object as returned by the processingChIPfunction
genomicProfiles	genomicProfiles object containing minimal information (such as the PWM)
parameters	vector or list containing each parameter that should be added to the chromosome. See generateStartingPopulation
generations	numeric describing the number of generation before the Genetic algorithm should halt.
mutationProbability	numeric descrbining the rate of mutations for each surviving individual
offsprings	numeric descrbining the number of individuals surviving to the next generation
chromatinState	GRanges object containing chromatin state information. Each state should be labled in a meta data column named "name". It is advised to use numeric values for each state name.
method	character string describing the scoring metric that should be used. ChIPanalyser offers twelve different metrics: correlation coefficients (Pearson, Spearman and Kendall), Mean Squared Error (MSE), Kolmogorov–Smirnov Distance, precision, recall, accuracy, F-score, Matthew’s correlation coefficient (MCC) and Area Under Curve Receiver Operator Characteristic (AUC ROC or just AUC)
lambda	logical describing if lambda value should be pre-computed. Setting to TRUE increases the speed of the algorithm.
checkpoint	logical describing if population parameters at each generations should be saved.
filename	character string that will serve as a prefix to the saved intermediate files.
cores	numeric describing the number of cores used to run the GA.

Details

ChIPanalyser offers a way of finding optimal solution by using a genetic algorithm. Instead of running the stadard analysis, TF binding affinities to chromatin states can be extracted via this more complex method. It should be noted that this method is better suited for the analysis of chromatin states. While the algorithm still works with simple DNA Accessibility, it would potentially take more time for accuracy minor gains.

Value

Returns a named list with three elements.

• database saves the data frame containing all scores for each individual since generation 1

• population saves the last population with chromosome values

• fitestsaves the fittest individual for a given generation

Author(s)

Patrick C.N. Martin <[e-mail]

Examples

library(ChIPanalyser)
data(ChIPanalyserData)
# See GA vignette for usage 

patrickCNMartin/ChIPanalyser documentation built on Nov. 24, 2022, 12:02 a.m.