RNA-Seq Analysis of Adipogenesis

Strand-specific paired-end libraries for RNA-Seq were generated from DNAse-treated total RNA using Ribozero and ScriptSeq systems (Epicentre/Illumina) and run on an Illumina 2500 sequencer to obtain 100 base paired-end reads. Low quality reads were filtered out by Kraken53 (link). The resulting filtered reads were mapped to the mouse genome version mm10 using Tophat54 (link). Mapped reads were counted with htseq-count55 (link) and read counts were normalized using deseq2 (ref. 56 (link)). In order to capture genes with the same expression pattern as the TAF and Oil Red O staining, said values were inserted into the normalized expression data set and then clustered with Biolayout express57 (link) using a 0.7 minimum Pearson correlation and a 95 correlation value. Clustering was conducted using MCL implementation of Markov Cluster Algorithm58 using an inflation coefficient of 2.2 and a preinflation coefficient of 3.0. The cluster of genes with included TAF and oil lipid data were extracted and analysed for GO over-representation using the PANTHER database59 (link).