The RMaP challenge of predicting RNA modifications by nanopore sequencing

Abstract

The field of epitranscriptomics is undergoing a technology-driven revolution. During past decades, RNA modifications like N6-methyladenosine (m⁶A), pseudouridine (ψ), and 5-methylcytosine (m⁵C) became acknowledged for playing critical roles in cellular processes. Direct RNA sequencing by Oxford Nanopore Technologies (ONT) enabled the detection of modifications in native RNA, by detecting noncanonical RNA nucleosides properties in raw data. Consequently, the field's cutting edge has a heavy component in computer science, opening new avenues of cooperation across the community, as exchanging data is as impactful as exchanging samples. Therefore, we seize the occasion to bring scientists together within the RNA Modification and Processing (RMaP) challenge to advance solutions for RNA modification detection and discuss ideas, problems and approaches. We show several computational methods to detect the most researched mRNA modifications (m⁶A, ψ, and m⁵C). Results demonstrate that a low prediction error and a high prediction accuracy can be achieved on these modifications across different approaches and algorithms. The RMaP challenge marks a substantial step towards improving algorithms' comparability, reliability, and consistency in RNA modification prediction. It points out the deficits in this young field that need to be addressed in further challenges.

Fig. 1. The RMaP challenge workflow.

RMaP challenge overview. a The several affiliations that contributed to the RMaP challenge. b Data preparation pipeline for each sub-challenge in RMaP. Datasets were prepared in vitro and measured with ONT. c General overview of the three sub-challenges proposed in RMaP. Each of them proposed a different task for selected RNA modifications. d The results obtained by the new methods are analyzed and compared.

Fig. 2. Method summary and results of RMaP challenge 1.

(Top) Summary of methods used for challenge 1. (Bottom) Comparison between Methods 1 and 2 performances on m⁵C modification detection. A lower value is better for all metrics. The diagram shows the values of each metric used in this work. The metrics values were obtained by comparing the two methods predictions with expected values. Metric values can be found in Table 2.

Fig. 3. Method summary and results of RMaP challenge 2.

(Top) Summary of methods used for challenge 2. (Bottom) Method 3 performance on m⁶A modification detection. A lower value is better for all metrics. The diagram shows the values of each metric used in this work. Metric values can be found in Table 2.

Fig. 4. Method summary and results of RMaP challenge 3.

(Top) Summary of methods used for challenge 3. (Bottom) Comparison between methods 4–7 performances on Ψ modification detection. The graph shows the values of each metric used in this work. The metrics values were obtained by comparing methods predictions with expected values. Metric values can be found in Table 2.

Fig. 5. Example bedRMod file.

Text visualization of a bedRMod file.