R/estimateScalingFactors-methods.R

In descostesn/ChIPSeqSpike: ChIP-Seq data scaling according to spike-in control

.computeBamCount <- function(bam_file, paired){
    
    result <- countBam(bam_file, 
            param=ScanBamParam(scanBamFlag(isPaired= paired, 
                            isUnmappedQuery =FALSE)))$records
    
    return(result)
}


.computeScalingFactor <- function(bam_count){
    
    return(1/(bam_count/1000000))
}


setMethod(
        
        f = "estimateScalingFactors",
        
        signature = "Experiment",
        
        definition = function(theObject, paired = FALSE, 
                verbose = TRUE){
            
            if(verbose) message("\t Computing endogenous scaling factor.")
            
            endo_count <- .computeBamCount(getBam(theObject), paired) 
            
            if(!endo_count) 
                stop("Endogenous bam file contains no aligned reads.")
            
            endo_SF <- .computeScalingFactor(endo_count)
            
            if(verbose) message("\t Computing exogenous scaling factor.")
            
            exo_count <- .computeBamCount(getExogenousBam(theObject),
                    paired)
            
            if(!exo_count) 
                stop("Exogenous bam file contains no aligned reads.")
            
            exo_SF <- .computeScalingFactor(exo_count)
            
            scalingFactor(theObject) <- endo_SF
            exogenousScalingFactor(theObject) <- exo_SF
            count(theObject) <- endo_count
            exoCount(theObject) <- exo_count
            
            return(theObject)
        })


setMethod(
        
        f = "estimateScalingFactors",
        
        signature = "ExperimentLoaded",
        
        definition = function(theObject, paired = FALSE, 
                verbose = TRUE){
            
            return(callNextMethod(theObject))
        })


setMethod(
        
        f = "estimateScalingFactors",
                
                signature = "ChIPSeqSpikeDataset",
                
                definition = function(theObject, paired = FALSE, 
                        verbose = TRUE){
                    
                    if(verbose) message("Computing scaling factor for input")
                    count(theObject) <- 
                            .computeBamCount(getBam(theObject), 
                                    paired)
                    
                    if(!getCount(theObject)) 
                        stop("input bam file contains no aligned reads.")
                    
                    scalingFactor(theObject) <- 
                            .computeScalingFactor(getCount(
                                            theObject))
                    
                    experimentList(theObject) <-
                            lapply(getExperimentList(theObject), 
                                    function(experiment){
                                        
                                        if(verbose)
                                            message("Processing ", 
                                                    getExpName(experiment))
                                        return(
                                 estimateScalingFactors(experiment, paired, 
                                         verbose))
                                    })
                    
                    return(theObject)
                })


setMethod(
        
        f = "estimateScalingFactors",
        
        signature = "ChIPSeqSpikeDatasetBoost",
        
        definition = function(theObject, paired = FALSE, 
                verbose = TRUE){
            
            if(verbose) message("Computing scaling factor for input")
            count(theObject) <- 
                    .computeBamCount(getBam(theObject), 
                            paired)
            
            if(!getCount(theObject)) 
                stop("input bam file contains no aligned reads.")
            
            scalingFactor(theObject) <- 
                    .computeScalingFactor(getCount(
                                    theObject))
            
            experimentList(theObject) <-
                    lapply(getExperimentList(theObject), 
                            function(experiment){
                                
                                if(verbose)
                                    message("Processing ", 
                                            getExpName(experiment))
                                return(
                                      estimateScalingFactors(experiment, 
                                              paired, verbose))
                            })
            
            return(theObject)
        })



.loadOnList <- function(object, paired, verbose){
    
    datasetList(object) <- lapply(getDatasetList(object), 
            function(object2){
                
                return(estimateScalingFactors(object2, paired, verbose)) 
            })
    
    return(object)
}

setMethod(f = "estimateScalingFactors",
        signature = "ChIPSeqSpikeDatasetList",
        definition = function(theObject, paired = FALSE, verbose = TRUE){
            .loadOnList(theObject, paired, verbose)
        })

setMethod(f = "estimateScalingFactors",
        signature = "ChIPSeqSpikeDatasetListBoost",
        definition = function(theObject, paired = FALSE, verbose = TRUE){
            .loadOnList(theObject, paired, verbose)
        })