Thank you for your comment. As design I consider every individual ("Line") a different condition, this because they all carry a different combination of the parental genomes, and each parent is considered one condition. We don't have 'real' different conditions like environmental stress. 

The code I use:

#read expression data

countDataGene <- as.matrix(read.csv("gene_count_matrix.csv", row.names="gene_id"))

#### Filter non expressed genes ####

#remove genes that are not very informative, remove genes which have <10 counts in 90% of samples

countDataGene1 <- countDataGene

m90 <- countDataGene1

percents_90 <- rowMeans(m90 > 10)

percents_90

m90_percents = m90[percents_90>0.1,]

head(m90_percents,100)

dim(m90_percents)

countDataGene <- m90_percents

#read in the experimental design

colData <- as.matrix(read.csv("Experdesign.csv", header= TRUE))

head(colData)

#all(rownames(Design_data) %in% colnames(countDataGene))

all(rownames(colData) %in% colnames(countDataGene))

#Create a DESeqDataSet from count matrix and labels

dds <- DESeqDataSetFromMatrix(countData = countDataGene, colData = colData, design = ~ Line)

dds

###### normalization #######

# Estimate size factors #

dds = estimateSizeFactors(dds)

#The correction (size) factors can be retrieved using:

sizeFactors(dds)

#estimate dispersions

dds <- estimateDispersions(dds)

plotDispEsts(dds)

######RLOG#######

rld <- rlog(dds, blind=FALSE)

rld_count_data <- assay(rld)

head(rld_count_data)

dim(rld_count_data)

#write.csv(rld_count_data, file ="Rlog_gene_count_data.csv", row.names = TRUE)

meanSdPlot(assay(rld))

plotDispEsts(dds)

#with top 30000 Transcripts with highest variance

topVarGenes <- head( order( rowVars( assay(rld) ), decreasing=TRUE ),30000)

topVarGenes <- assay(rld)[ topVarGenes, ]