# Input: 'E' is a bulk expression matrix (rows: genes, cols: samples),

# 'C' is a cell type proportion matrix (rows: cell types, cols: samples)

# Output: This function checks that all inputs are valid and returns them in controlled format

CheckInput = function(E, C){

# Check that 'E' and 'C' are some sort of tables and convert them into matrices

if(!any( class(E) %in% c("matrix", "data.frame") )) stop("'E' should be a matrix or a data frame.")

if(!any( class(C) %in% c("matrix", "data.frame") )) stop("'C' should be a matrix or a data frame.")

E = as.matrix(E)

C = as.matrix(C)

# 'E' and 'C' should have the same number of samples

if(ncol(C) != ncol(E)) stop("Bulk matrix 'E' and cell type proportion matrix 'C' should have equal number of columns (i.e. samples).")

# Remove genes that are unexressed in at least 40% of samples (in both sample groups)

# if(is.na(groups)){

# limit = floor(0.4*ncol(E))

# dropgenes = which(apply(E, 1, function(g){sum(g==0)}) >= limit)

# #if(length(dropgenes) > 0) E = E[-dropgenes,]

# } else{

# limit1 = floor(0.4*sum(groups==1))

# limit2 = floor(0.4*sum(groups==2))

# dropgenes1 = which(apply(E[,groups==1], 1, function(g){sum(g==0)}) >= limit1)

# dropgenes2 = which(apply(E[,groups==2], 1, function(g){sum(g==0)}) >= limit2)

# dropgenes = intersect(dropgenes1, dropgenes2)

# if(length(dropgenes) > 0) E = E[-dropgenes,]

# }

return(list(E=E, C=C))

}