*4.10. Bioinformatics Analysis*

Using random primers, the Lexogen Quantseq kit (Lexogen) generates fragments that end with a poly A tail. Since this random primer may introduce mistakes, the first 11bp of all reads were trimmed. Poly A tails and adapters at the end of each read were trimmed, along with basic quality trimming. Only reads longer than 20 bps were kept for further processing. All trimming and filtering steps were done using bbduk of bbtools 36.84 [99]. Reads were mapped to the Human genome (hg38) using STAR 2.5 [100]. Bam file modifications were done using elprep 2.5 [101]. Htseq 0.6.1p1 [102] was used to count the number of mapped reads per known gene. The gene definitions of Ensembl 87 were used. Reads were only considered in the counting process if the mapping quality was equal to or higher than 10, the strand of the read was the same strand as the gene and the read was not mapped in overlapping gene definitions (the union option). Differential expression analysis was performed using tools DESeq2 [103] and EdgeR [104,105]. The results of these tools were merged. Only genes that were significant with an FDR <0.1 in both tools were considered for further pathway analysis.
