Introduction: Alternative splicing is an important process in the heart to maintain proteomic diversity and function. We aimed to map the transcriptional changes in cardiomyopathies compared to healthy controls and to uncover thousands of novel isoforms of important cardiac transcripts.
Methods: Short-read Illumina sequencing was complemented by Nanopore long-read sequencing using the GridION device. Full-length cDNA libraries were generated from human cardiac tissue and after introduction of synthetic spike-ins sequencing was performed on in total 30 samples.
Results: The mean sequencing quality resulted in a phred score of 8.6 ± 0.5 (Mean ± STD). From all generated reads, 72 ± 16 % (Mean ± STD) aligned to the human genome (GRCh38). A total of 26 ± 17 % are assigned to the synthetic spike-ins and only 3 ± 1 % (Mean ± STD) were unmapped.  The produced median read length was 691 ± 82 bp (Mean ± STD). In total, our approach produced 9.6 ± 3.9 million reads (Mean ± STD) of which 48 ± 8 % (Mean ± STD) could be classified as full-length reads. Analysis on gene- and transcriptome level showed clustering of the disease vs. healthy groups. Analyzing the differentially expressed isoforms, we were able to detect around 20,400 novel isoforms. With high significance (p-value < 0.001) we could find previously unknown candidate isoforms that underline a functional reassembly of the cytoskeleton and sarcomere. These results could be independently replicated using isoform-specific quantitative PCR.
Conclusion: Long-read sequencing is able to map the cardiac transcriptome beyond gene expression changes and highlights the important role of isoform switching. The results are the basis to uncover novel pathological mechanisms.