Nanopore-based native RNA sequencing provides insights into prokaryotic transcription, operon structures, rRNA maturation and modifications

Felix Grünberger; Robert Knüppel; Michael Jüttner; Martin Fenk; Andreas Borst; Robert Reichelt; Winfried Hausner; Jörg Soppa; Sebastien Ferreira-Cerca; Dina Grohmann

we detected a decrease in coverage from the 3´ to 5´end of the RNA in all RNA classes except for 465 the spike-in control (see Supplementary Fig. 9b ), which is a limitation reported in the 466 literature 13,60,61 . Therefore, we assume that not Nanopore sequencing but library preparation 467 causes this problem. Based on this information, we developed a strategy that first collapses all 468 overlapping reads and then splits them according to a significant coverage drop on the 3´ ends 469 (annotation of TUs based on this strategy in Supplementary Table 5) . We compared the results to 470 database annotations and found that most of the differences are either caused by the low 471 sequencing depth or by single-unit operons that have been collapsed and are now two-unit 472 operons in the ONT data sets (see Supplementary Fig. 10a,b) 29,50 . Even though limited read 473 availability is a concern in all data sets, many large operons were detected for all organisms (see 474 Supplementary Fig. 10c ). In case of limited bioinformatical resources, TUs can be explored visually 475 in a genome browser, which is mostly not possible for Illumina reads (Fig. 3 , see Supplementary 476 Fig. 11,12) . It further allows a quantitative analysis of individual transcripts in relation to other 477 elements of the TU and performs much better than pure bioinformatical prediction or molecular 478 biology methods (RT-PCR) as shown for the archaellum operon in P. furiosus (Fig. 3) seq data, the analysis shows that this operon is split into two transcription units. This shows that 484 the ONT native RNA sequencing method provides the opportunity to correctly annotate 485 transcriptional units thereby outperforming the bioinformatics-only prediction as well as the 486 visual inspection of Illumina coverage (see Supplementary Fig. 11) . Besides, we confirmed the 487 complex transcription pattern of the major ribosomal protein gene cluster in E. coli that stretches 488 over more than 10 kB, including the accurate determination of TSS and TTS and a putative cleavage 489 site in the secY gene (see Supplementary Fig. 12) 63 . 490 the author/funder. All rights reserved. No reuse allowed without permission.