Nanopore sequencing of RNA and cDNA molecules in Escherichia coli

Abstract

High-throughput sequencing dramatically changed our view of transcriptome architectures and allowed for ground-breaking discoveries in RNA biology. Recently, sequencing of full-length transcripts based on the single-molecule sequencing platform from Oxford Nanopore Technologies (ONT) was introduced and is widely employed to sequence eukaryotic and viral RNAs. However, experimental approaches implementing this technique for prokaryotic transcriptomes remain scarce. Here, we present an experimental and bioinformatic workflow for ONT RNA-seq in the bacterial model organism Escherichia coli, which can be applied to any microorganism. Our study highlights critical steps of library preparation and computational analysis and compares the results to gold standards in the field. Furthermore, we comprehensively evaluate the applicability and advantages of different ONT-based RNA sequencing protocols, including direct RNA, direct cDNA, and PCR-cDNA. We find that (PCR)-cDNA-seq offers improved yield and accuracy compared to direct RNA sequencing. Notably, (PCR)-cDNA-seq is suitable for quantitative measurements and can be readily used for simultaneous and accurate detection of transcript 5'and 3' boundaries, analysis of transcriptional units and transcriptional heterogeneity. In summary, based on our comprehensive study, we show that Nanopore RNA-seq to be a ready-to-use tool allowing rapid, cost-effective, and accurate annotation of multiple transcriptomic features. Thereby Nanopore RNA-seq holds the potential to become a valuable alternative method for RNA analysis in prokaryotes.

Development of ONT-cappable-seq to unravel the transcriptional landscape of Pseudomonas phages

“…In recent years, several full-length RNA-seq protocols have emerged to comprehensively characterize prokaryotic transcripts (14)(15)(16). Notably, the SMRT-Cappable-seq protocol includes size-selection steps to eliminate all cDNA fragments below 1 kb from the final library (15).…”

“…Next, the raw reads were processed by Pychopper (v2.5.0) (https://github.com/nanoporetech/pychopper) to identify, rescue and correctly orient full-length cDNA reads. Cutadapt (v2.7) (32) was used to remove 3’ polyA stretches and sequence remnants on the 5’ end that match the primers used during reverse transcription, as described previously (16).…”

“…The identification of transcriptional boundaries was carried out using custom scripts inspired by previously published work (16), with modifications tailored to our ONT-cappable-seq approach (https://github.com/LoGT-KULeuven/ONT-cappable-seq). For TSSs detection, the alignment files were used to generate strand-specific bed files that report the number of reads starting at each position in the LUZ7 genome using bedtools genomecov (−5 -bga) (37).…”

“…For TSSs detection, the alignment files were used to generate strand-specific bed files that report the number of reads starting at each position in the LUZ7 genome using bedtools genomecov (−5 -bga) (37). Next, a peak calling algorithm (https://github.com/NICHD-BSPC/termseq-peaks) was applied to determine genomic positions of local maxima from 5’ read ends (16, 38). Nearby peaks within a distance of 5 bp on the same strand were clustered, retaining only the peak position with the highest number of 5’ read ends.…”

“…However, as this method uses size selection to enrich for longer transcripts, it is less suited for the exploration of phage transcriptional landscapes, which are often hallmarked by high gene densities and short open reading frames. In parallel, Oxford Nanopore's long-read sequencing technology (ONT) is rapidly improving in quality and is emerging as a cost-effective technique in microbial transcriptomics (12,(16)(17)(18).…”