gaucho_vignette.R
In gaucho: Genetic Algorithms for Understanding Clonal Heterogeneity and Ordering

## ----load_library_fake, eval=FALSE---------------------------------------
#  ## Install
#  source("http://bioconductor.org/biocLite.R")
#  biocLite("gaucho")
#  ## Load
#  library(gaucho)

## ----demontrate_proportions, eval=TRUE-----------------------------------
## First sample is reprented by first block of three digits
## First sample is 100% clone 1, 0% clone 2, 0% clone 3
c(5,0,0)/sum(c(5,0,0))
## Second sample is reprented by second block of three digits
## Second sample is 30% clone 1, 70% clone 2, 0% clone 3
c(3,7,0)/sum(c(3,7,0))
## Third sample is reprented by third block of three digits
## Third sample is 0% clone 1, 20% clone 2, 80% clone 3
c(0,1,4)/sum(c(0,1,4))

## ----emit_results, eval=FALSE--------------------------------------------
#  ## Write results to files
#  gauchoReport(gaucho_input_dataframe,gaucho_output_object)

## ----gaucho_simple_data_show, eval=TRUE----------------------------------
## Load library
library(gaucho)

## Load simple data set
gaucho_simple_data = read.table(file.path(system.file("extdata",package="gaucho"),"gaucho_simple_data.txt"),header=TRUE,row.names=1)

## There  are three columns (time points T0, T1 and T2)
## and three rows (mutations M1, M2, M3)
gaucho_simple_data

## ----gaucho_simple_data_execute, eval=FALSE------------------------------
#  ## Execute gaucho() function on the simple data set
#  simpleDataSolution=gaucho(gaucho_simple_data, number_of_clones=3, nroot=1,iterations=1000)

## ----gaucho_simple_data_report, eval=FALSE-------------------------------
#  ## Produce plots for the phylogeny, heatmap and proportions in turn
#  gauchoReport(gaucho_simple_data,simpleDataSolution,outType="phylogeny")
#  gauchoReport(gaucho_simple_data,simpleDataSolution,outType="heatmap")
#  gauchoReport(gaucho_simple_data,simpleDataSolution,outType="proportion")
#  
#  ## Create output files representing the solution(s) in the current working directory
#  gauchoReport(gaucho_simple_data,simpleDataSolution,outType="complete")
#  
#  ## In case you want to know the current working directory, it can be reported using this function:
#  getwd()

## ----gaucho_hidden_data_load, eval=TRUE----------------------------------
## Load library
library(gaucho)

## Load hidden data set
gaucho_hidden_data = read.table(file.path(system.file("extdata",package="gaucho"),"gaucho_hidden_data.txt"),header=TRUE,row.names=1)

## There  are three columns (time points T0, T1 and T2)
## and five rows (mutations M1, M2, M3, M4 and M5)
gaucho_hidden_data

## ----gaucho_hidden_data_results, eval=FALSE------------------------------
#  ## Produce plots for the phylogeny, heatmap and proportions in turn
#  gauchoReport(gaucho_hidden_data,hiddenDataSolution,outType="phylogeny")
#  gauchoReport(gaucho_hidden_data,hiddenDataSolution,outType="heatmap")
#  gauchoReport(gaucho_hidden_data,hiddenDataSolution,outType="proportion")
#  
#  ## Create output files representing the solution(s) in the current working directory
#  gauchoReport(gaucho_simple_data,simpleDataSolution,outType="complete")

## ----gaucho_synthetic_data, eval=FALSE-----------------------------------
#  ## Load library
#  library(gaucho)
#  
#  ## Load synthetic data set and the same data with added jitter
#  gaucho_synth_data = read.table(file.path(system.file("extdata",package="gaucho"),"gaucho_synth_data.txt"),header=TRUE,row.names=1)
#  gaucho_synth_data_jittered = read.table(file.path(system.file("extdata",package="gaucho"),"gaucho_synth_data_jittered.txt"),header=TRUE,row.names=1)
#  
#  ## Execute gaucho() function on the synthetic data set - we know that
#  ## there are 6 clones
#  s=gaucho(gaucho_synth_data,number_of_clones=6,iterations=3000)
#  
#  ## Access solution slot in returned object to show highest scoring solution(s)
#  ## The optimum solution's score is -2.22E-15, which is a rounding error from zero
#  s@solution
#  
#  ## Create output files representing the solution(s)
#  ## in the current working directory
#  gauchoReport(gaucho_synth_data,s)

## ----gaucho_yeast_data, eval=FALSE---------------------------------------
#  ## Load library
#  library(gaucho)
#  
#  ## Load yeast data set
#  BYB1_G07_pruned = read.table(file.path(system.file("extdata",package="gaucho"),"BYB1_G07_pruned.txt"),header=TRUE,row.names=1)
#  
#  ## Execute gaucho() function on the yeast data set
#  ## The paper claims that there are 6 clones with multiple roots
#  yDataSolution=gaucho(BYB1_G07_pruned, number_of_clones=6, contamination=1, iterations=3000)
#  
#  ## Access solution slot in returned object to show highest scoring solution(s)
#  yDataSolution@solution
#  
#  ## Create output files representing the solution(s) in the current working directory
#  gauchoReport(BYB1_G07_pruned,yDataSolution)

## ----gaucho_find_clones, eval=FALSE--------------------------------------
#  ## Load library
#  library(gaucho)
#  
#  ## Load simple data set
#  gaucho_synth_data = read.table(file.path(system.file("extdata",package="gaucho"),"gaucho_synth_data.txt"),header=TRUE,row.names=1)
#  
#  ## We know that there are 6 clones in this data set.
#  ## Let's loop from number_of_clones=4 to number_of_clones=8 to illustrate this
#  ## Also, assume we know nothing about phylogeny, so leave nroot as the default value
#  
#  ## Assign parameters and create an empty data frame to hold the results
#  clonerange=4:8
#  n=5
#  results=matrix(NA,nrow=iterations,ncol=length(clonerange))
#  colnames(results)=paste0(rep("clone",length(clonerange)),clonerange)
#  
#  ## Execute gaucho n times for each number of clones
#  for(c in clonerange){
#    for(thisn in 1:n){
#      message(paste("Iteration",thisn,"using",c,"clones"))
#      s=gaucho(gaucho_synth_data, number_of_clones=c,iterations=1000)
#      ## Assign best fitness value to an element in the matrix
#      results[thisn,which(c==clonerange)]=s@fitnessValue
#    }
#  }
#  
#  ## Plot the results
#  boxplot(results)

## ----gaucho_multi_runs, eval=FALSE---------------------------------------
#  ## Load library
#  library(gaucho)
#  
#  ## Load simple data set
#  gaucho_synth_data = read.table(file.path(system.file("extdata",package="gaucho"),"gaucho_synth_data.txt"),header=TRUE,row.names=1)
#  
#  ## Execute 20 runs
#  n=20
#  
#  ## Create an empty list to hold all results
#  gauchoResults = list()
#  
#  ## Execute gaucho n times, pushing each resulting object into a list
#  for(thisn in 1:n){
#    message(paste("Iteration",thisn))
#    syntheticDataSolution=gaucho(gaucho_synthetic_data, number_of_clones=6, iterations=10000)
#    gauchoResults[thisn]=syntheticDataSolution
#  }
#  
#  ## Find lowest fitness values of all runs
#  bestScore = max(sapply(gauchoResults, function(x) x@fitnessValue))
#  
#  ## Which run was it?
#  which(sapply(gauchoResults, function(x) x@fitnessValue)==bestScore)

## ----sessionInfo, eval=TRUE----------------------------------------------
sessionInfo()