DiffBUM-HMM: a robust statistical modeling approach for detecting RNA flexibility changes in high-throughput structure probing data

Ka Ying Toby Law,Paolo Marangio,Sander Granneman,Guido Sanguinetti

doi:10.1186/s13059-021-02379-y

Ka Ying Toby Law, Paolo Marangio + Show 2 more

Open Access

https://doi.org/10.1186/s13059-021-02379-y

Copy DOI

Abstract

Advancing RNA structural probing techniques with next-generation sequencing has generated demands for complementary computational tools to robustly extract RNA structural information amidst sampling noise and variability. We present diffBUM-HMM, a noise-aware model that enables accurate detection of RNA flexibility and conformational changes from high-throughput RNA structure-probing data. diffBUM-HMM is widely compatible, accounting for sampling variation and sequence coverage biases, and displays higher sensitivity than existing methods while robust against false positives. Our analyses of datasets generated with a variety of RNA probing chemistries demonstrate the value of diffBUM-HMM for quantitatively detecting RNA structural changes and RNA-binding protein binding sites.

Highlights

Understanding the structure of RNA is key to unravel its in vivo function, and it is highly relevant to biomedicine, drug discovery, and synthetic biology [1,2,3,4]
For diffBUM-HMM, the hidden state is expanded to take on four potential values instead of two: nucleotide is unmodified in both conditions (UU; hidden state 1); nucleotide is unmodified in the 1st condition but modified in the 2nd (UM; hidden state 2); nucleotide is modified in the 1st condition, but unmodified in the 2nd (MU; hidden state 3); nucleotide is modified in both conditions (MM; hidden state 4)
We found that in the ex vivo data for both diffBUM-HMM and deltaSHAPE many of the RBP binding sites were statistically significantly enriched for differentially reactive nucleotides (DRNs), with diffBUMHMM DRNs preferentially enriched in CELF1 and FUS binding sites (Fig. 5D)

Summary

Introduction

Understanding the structure of RNA is key to unravel its in vivo function, and it is highly relevant to biomedicine, drug discovery, and synthetic biology [1,2,3,4]. Recent years have witnessed a blossoming of high-throughput methods that couple nextgeneration sequencing with biochemical assays to ‘probe’ the structure of thousands of RNA molecules simultaneously, including whole transcriptomes [5,6,7,8,9,10,11,12,13,14,15,16] The majority of these biochemical assays use reagents such as SHAPE (selective 2 -hydroxyl acylation analyzed by primer extension) reagents [5,6,7, 16,17,18,19,20] and dimethyl sulfate (DMS) [8, 10, 21]. It is not immediately usable for differential analyses between different treatments

Methods

Results

Discussion

Conclusion