Robust statistical modeling greatly improves sensitivity of high-throughput RNA structure probing experiments

Structure probing coupled with high-throughput sequencing holds the potential to revolutionise our understanding of the role of RNA structure in regulation of gene expression. Despite major technological advances, intrinsic noise and high coverage requirements greatly limit the applicability of these techniques. Existing methods [1, 2, 3] do not provide strategies for correcting biases of the technology and are not sufficiently informed by inter-replicate variability in order to perform justifiable statistical assessments.
We developed a probabilistic modelling pipeline which specifically accounts for biological variability and provides automated empirical strategies to correct coverage- and sequence-dependent biases in the data. The output of our method yields statistically interpretable scores for the probability of nucleotide modification transcriptome-wide, obviating the need for arbitrary thresholds and post-processing. We demonstrate on two yeast data sets that our method has greatly increased sensitivity, enabling the identification of modified regions on a greatly increased number of transcripts, compared with existing pipelines. Our method also provides accurate and confident predictions at much lower coverage levels than those recommended in recent studies [3, 4], which are normally only met for a handful of transcripts in transcriptome-wide experiments. Our results show that statistical modelling greatly extends the scope and potential of transcriptome-wide structure probing experiments."

[1] Ding, Yiliang, et al. Nature 505.7485 (2014).
[2] Kielpinski, Lukasz Jan, and Jeppe Vinther, Nucleic acids research (2014).
[3] Talkish, Jason, et al. RNA 20.5 (2014).
[4] Siegfried, Nathan A., et al. Nature methods 11.9 (2014).