Background:
Oxford Nanopore’s direct RNA sequencing (DRS) is a powerful tool capable of sequencing complete RNA molecules and accurately measuring both gene and isoform expression levels. However, as DRS is designed to profile intact native RNA the precision and reliability of gene and isoform quantification may be more heavily dependent upon RNA integrity, (as measured by RIN), than other RNA-seq methodologies. Currently it is unclear what range of RNA integrity DRS can be applied to and if the impact of degradation can be corrected for. RNA from in-vivo or clinical samples have often undergone some degradation, therefore, to allow application of DRS to the widest range of suitable samples it is essential to determine the appropriate range of RIN for DRS.
Results:
To determine the impact of RNA degradation on DRS we performed a degradation timecourse, sequencing SHSY5Y neuroblastoma samples where the RNA was allowed to degrade prior to RNA extraction. Our data demonstrate that degradation is a significant and pervasive factor that can bias downstream analysis. This includes a reduction in library complexity resulting in an overrepresentation of short genes and isoforms. Increased gene and isoform length were also closely correlated with rate of degradation, with DNA binding proteins and transcription factors showing the most degradation susceptibility. Degradation also biases differential expression analyses; however, we find explicit correction can almost fully recover meaningful biological signal. Overall, we find samples with RIN greater that 9.6 can be treated as undegraded and samples with RIN greater than 7 can be utilised for DRS with appropriate sample correction.
Conclusions:
These results serve to inform the use of DRS by providing a framework for modelling samples impacted by degradation. In so doing DRS can be appropriately applied across a variety of sample RINs whilst limiting the confounding effect of degradation on downstream analysis.