SRA

3' RNA-Seq and direct RNA hybridisation are conceptually different approaches for high throughput transcriptome analysis. The former generates next-generation sequencing libraries by targeting and amplifying 3'-end sequences of 60 to 80 nucleotides in length, whereas the latter quantifies transcripts of 800 pre-selected genes. Since both methods are applicable to short mRNA fragments, they are thought to allow for transcriptome analyses from formalin-fixed, paraffin-embedded (FFPE) archival tissue containing partially degraded RNA. In this study, results of both approaches were compared on sample- and gene-wise count levels, gene expression strength and direction, as well as the overlap of differentially expressed genes (DEGs). Using two different oncological perspectives, i.e. a stage-dependent and entity-contrasting comparison on 35 FFPE canine tumours, both methods proved suitable for their use on archival tissues. A moderately to very strong overall count correlation was found (range of Pearson and Spearman means: 0.66 – 0.87). Of note, the gene-wise count correlations depended on gene expression strength. Finally, in the entity-contrasting comparison, expression direction correlated very strongly (range: 0.88 – 0.91) but DEGs only moderately overlapped (Jaccard index: 0.53). Aside from these figures, different practically relevant aspects of the two technologies offer distinct advantages, depending on the objectives and design of a study. - Attention: this GEO submission includes the 3' RNA-Seq data from 25 of the 35 samples used in this study! The data from the RNA hybridisation platform (nCounter) and the other 10 stage-contrasting tumour samples were submitted separately to GEO. Overall design: To determine the correlation of gene expression data generated by the nCounter® Canine IO Panel and QuantSeq 3' from three types of canine tumours. FFPE tissue samples were obtained from the archive (spanning the years 2013 to 2021) and total RNA was extracted. After QuantSeq 3' and hybridisation to the nCounter® Canine IO Panel with bioinformatic processing, counts and differential gene expression data were compared to calculate correlations between the two techologies employed. - Attention: The data in this GEO submission comes from the entity-contrasting comparison of the two most common glandular, perianal tumours in the dog – the hepatoid gland adenoma (HGA, n = 10) and apocrine gland anal sac adenocarcinoma (AGASAC, n = 15). The other stage-contrasting tumour samples (n = 10) used in the publication were submitted separately to GEO.