RNA modifications detection by comparative Nanopore direct RNA sequencing

Abstract

RNA molecules undergo a vast array of chemical post-transcriptional modifications (PTMs) that can affect their structure and interaction properties. In recent years, a growing number of PTMs have been successfully mapped to the transcriptome using experimental approaches relying on high-throughput sequencing. Oxford Nanopore direct-RNA sequencing has been shown to be sensitive to RNA modifications. We developed and validated Nanocompore, a robust analytical framework that identifies modifications from these data. Our strategy compares an RNA sample of interest against a non-modified control sample, not requiring a training set and allowing the use of replicates. We show that Nanocompore can detect different RNA modifications with position accuracy in vitro, and we apply it to profile m6A in vivo in yeast and human RNAs, as well as in targeted non-coding RNAs. We confirm our results with orthogonal methods and provide novel insights on the co-occurrence of multiple modified residues on individual RNA molecules.

Prediction of m6A and m5C at single-molecule resolution reveals a cooccurrence of RNA modifications across the transcriptome

“…We next compared CHEUI-solo with Nanocompore (Leger et al 2021), Xpore (Pratanwanich et al 2021), and Epinano (Liu et al 2019) for their ability to detect and quantify RNA modifications using DRS reads. To achieve this, we built positive and negative independent test datasets.…”

“…The first one includes methods that rely on comparing two conditions, one corresponding to a sample of interest, often the wild type (WT) sample, and the other with a reduced or abolished presence of a specific modification, usually obtained through a knock-out (KO) or knock-down (KD) of a modification ‘writer’ enzyme. This category includes Nanocompore (Leger et al 2021), Xpore (Pratanwanich et al 2021), DRUMMER (Price et al 2020), nanoDOC (Ueda 2020), Yanocomp (Parker et al 2021) and Tombo in sample comparison mode. These methods compare collective properties of DRS signals in the two conditions.…”

The S. cerevisiae m6A-reader Pho92 impacts meiotic recombination by controlling key methylated transcripts

“…Large sets of m 6 A sites had been previously reported in yeast mRNAs during meiosis (Bushkin et al, 2019; Dierks et al, 2021; Garcia-Campos et al, 2019; Leger et al, 2021; Schwartz et al, 2013). Those were established using different methods and show limited overlap, partly owing to distinct experimental parameters (different strains or analyses at different time points) but also possibly due to methodological limitations (McIntyre et al, 2020).…”

Concerted modification of nucleotides at functional centers of the ribosome revealed by single-molecule RNA modification profiling

“…While there is some evidence that unknown modified kmer distributions can be estimated using known kmer distributions (Ding et al, 2021), generating more specific modification training data sets that contain all combinations of partially modified closely spaced clusters of nucleotides may be required to produce more accurate and general modification detection algorithms. This is especially true if de-novo detection of modifications within complex sequences is the goal (Leger et al, 2019; Stoiber et al, 2017).…”

“…However, if a canonical position is directly next to a modified position, it is very likely the underlying current is going to be shifted for that position. Also, the uncertainty of which specific nucleotide in the pore gives rise to the most significant signal shift makes site selection for kmer based sample compare frameworks very difficult (Ding et al, 2020; Leger et al, 2019; Stoiber et al, 2017) . Therefore, instead of evaluating Tombo on the per-position modification calling accuracy, we used a less stringent metric of modification window calling accuracy.…”