#Load miodin
library(miodin)

#Create project
mp <- MiodinProject(name = "VerticalIntegration")

#Study design
library(data.table)
pheno <- fread(system.file("extdata/tcgaPheno.txt", package = "miodindata"))  #Read pheno data from miodindata package
sampleTable <- data.frame(
  SampleName = pheno$patientID,                                      #Sample names
  SamplingPoint = rep("sp1", nrow(pheno)),                           #One sampling point (must be specified)
  VitalStatus = ifelse(pheno$vital_status == 1, "Alive", "Diseased") #Whether samples are alive or diseased
)
assayTableRNA <- data.frame(
  SampleName = pheno$patientID,                                      #Sample names
  DataFile = rep(system.file("extdata/tcgaRNA.txt",                  #Data file containing mRNA expression
                             package = "miodindata"), nrow(pheno)),
  DataColumn = pheno$patientID                                       #Column in data file corresponding to each sample
)
assayTableMiRNA <- data.frame(
  SampleName = pheno$patientID,                                      #Sample names
  DataFile = rep(system.file("extdata/tcgaMiRNA.txt",                #Data file containing miRNA expression
                             package = "miodindata"), nrow(pheno)),
  DataColumn = pheno$patientID                                       #Column in data file corresponding to each sample
)
assayTableMethyl <- data.frame(
  SampleName = pheno$patientID,                                      #Sample names
  DataFile = rep(system.file("extdata/tcgaMethyl.txt",               #Data file containing methylation values
                             package = "miodindata"), nrow(pheno)),
  DataColumn = pheno$patientID                                       #Column in data file corresponding to each sample
)
ms <- studyDesignCaseControl(                      #Create a miodin stiudy (case-control design)
  studyName = "TCGAStudy",                         #Name of study in project (must be unique)
  factorName = "VitalStatus",                      #Column in sample table containing vital status of patients
  caseName = "Diseased",                           #Factor level corresponding to cases
  controlName = "Alive",                           #Factor level corresponding to controls
  contrastName = "VitalStatus",                    #Name of contrast representing the comparison between cases and controls
  numCase = 287,                                   #Number of cases
  numControl = 51,                                 #Number of controls
  sampleTable = sampleTable,                       #Sample table data frame
  assayTable = assayTableRNA,                      #Assay table data frame (mRNA)
  assayTableName = "RNAseq") +                     #Assay table name (mRNA)
  studyAssayTable("miRNAseq", assayTableMiRNA) +   #Adding another assay table (miRNA)
  studyAssayTable("Methyl", assayTableMethyl)      #Adding another assay table (methylation)
insert(ms, mp)                                     #Insert study into project
rm(pheno, sampleTable, assayTableRNA, assayTableMiRNA, assayTableMethyl, ms)

#Import and process
mw <- MiodinWorkflow("TCGAFlow")  #Create a new workflow
mw <- mw +
  importProcessedData(                 #Module to import processed (counts) RNA-seq data
    name = "RNA importer",             #Name of module in workflow (must be unique)
    experiment = "sequencing",         #Type of experimental platform (microarray, sequencing or mass spec)
    dataType = "rna",                  #Omics data type (rna, snp, mutation or methylation or protein)
    studyName = "TCGAStudy",           #Name of study in project
    assayName = "RNAseq",              #Name of assay table in study
    datasetName = "TCGAData",          #Name of output dataset
    contrastName = "VitalStatus"       #Name of contrast to consider for coloring samples in QC plots
  ) %>%
  processSequencingData(               #Module to process imported RNA-seq data
    name = "RNA processor",            #Name of module in workflow
    contrastName = "VitalStatus",      #Name of contrast (determines which samples to process)
    filterLowCount = TRUE,             #Filter genes with low counts
    stabilizeVariance = TRUE,          #Perform variance stabilization
    stabilizeVarianceMethod = "vst"    #Normalization method for variance stabilization
  ) +
  importProcessedData(                 #Import miRNA-seq count data
    name = "miRNA importer",
    experiment = "sequencing",
    dataType = "rna",
    studyName = "TCGAStudy",
    assayName = "miRNAseq",
    datasetName = "TCGAData",
    contrastName = "VitalStatus"
  ) %>%
  processSequencingData(               #Process imported miRNA-seq data
    name = "miRNA processor",
    contrastName = "VitalStatus",
    filterLowCount = TRUE,
    stabilizeVariance = TRUE,
    stabilizeVarianceMethod = "vst"
  ) +
  importProcessedData(                 #Import methylation data
    name = "Methyl importer",
    experiment = "microarray",
    dataType = "methylation",
    studyName = "TCGAStudy",
    assayName = "Methyl",
    datasetName = "TCGAData",
    contrastName = "VitalStatus",
    chipName = "HumanMethylation450K"  #Name of microarray used to measure methylation
  ) %>%
  processMicroarrayData(               #Process imported methylation data
    name = "Methyl processor",
    contrastName = "VitalStatus",
    filterSNPandCH = TRUE,             #Filter out bad probes (according to DMRcate package)
    filterNonCG = TRUE                 #Filter out non-CG probes
  )
mw <- insert(mw, mp)  #Insert workflow into project
mw <- execute(mw)     #Execute workflow

#Integrate and analyze
mw <- mw +
  integrateAssays(                            #Module to integrate data into single dataset
    name = "Assay integrator",                #Name of module
    upstreamModules = c("RNA processor",      #Upstream modules from which data should be integrated
                        "miRNA processor",
                        "Methyl processor"),
    datasetName = "TCGAData"                  #Name of output dataset
  ) %>%
  performFactorAnalysis(                      #Module for performing factor analysis
    name = "Factor analyzer",                 #Name of module
    analysisMethod = "mofa",                  #Analysis method (mofa or bsplsda)
    annotColor = "VitalStatus",               #Contrast to use for coloring samples in plots
    numFactors = 2,                           #Number of latent factors in the model
    seed = 123                                #Set seed to render results reproducible
  )
mw <- execute(mw)

#Export results
export(mp, "dataset", "TCGAData")             #Export TCGAData dataset to disk
export(mp, "analysisResult", "TCGAData")      #Export analysis result to disk
saveDataFile(mp)                              #Save project and objects in hierarchy