R/getSumreplicateschild.R

In jaspershen/MSannotator: xMSannotator: R package for network based annotation of metabolomics data.

getSumreplicateschild <-
function(curdata,alignment.tool,numreplicates,rep.max.missing.thresh,method="mean",missing.val)
{
       #curdata<-t(curdata)
       #write.table(curdata,file="test.txt",sep="\t",row.names=FALSE)
        numfeats=dim(curdata)[1]
        numsamp=dim(curdata)[2]
     # if(FALSE){
        resvec_1<-lapply(1:numfeats,function(r)
        {
                newrow={}
                finalmat={}
                #for(samp in seq(1,(numsamp),numreplicates))
                {
                       # i=samp
                        #j=i+numreplicates-1

                        curdata_int=curdata[r,]
			
						if(is.na(missing.val)==FALSE){
                        check_zeros=which(curdata_int==missing.val)
                        }else{
                        	check_zeros=which(is.na(curdata_int)==TRUE)
                        	}
                        na_thresh=round(rep.max.missing.thresh*numreplicates)
                        
                        
                        if(length(check_zeros)>na_thresh)
                        {
                                meanval<-missing.val
                        }
                        else
                        {
                                #temporarily replace the missing intensities, set to 0 in apLCMS,
                                #with mean intensity value of the corresponding replicates (with non-zero values)
                                #curdata_int[check_zeros]=mean(t(curdata_int[-c(check_zeros)]))
                                 if(length(check_zeros)>0)
                                {
                                		if(method=="mean"){
                                			 meanval<-mean(t(curdata_int[-check_zeros]),na.rm=TRUE)
                                		}else{
                                        meanval<-median(t(curdata_int[-check_zeros]),na.rm=TRUE)
                                		}
                                }
                                else
                                {
                                		if(method=="mean"){
                                				meanval<-mean(t(curdata_int),na.rm=TRUE)
                                			}else{
                                        meanval<-median(t(curdata_int),na.rm=TRUE)
                                		}
                                }

                        }
                        newrow<-cbind(newrow,meanval)
                }


                finalmat<-rbind(finalmat, newrow)
                return(finalmat)
        })
        
        #colnames(final_set)<-colnames_data
        #rownames(final_set)=NULL
        return(resvec_1)
	
	
	
}