Analysis of Ribosome Stalling and Translation Elongation Dynamics by Deep Learning

Abstract

Ribosome stalling is manifested by the local accumulation of ribosomes at specific codon positions of mRNAs. Here, we present ROSE, a deep learning framework to analyze high-throughput ribosome profiling data and estimate the probability of a ribosome stalling event occurring at each genomic location. Extensive validation tests on independent data demonstrated that ROSE possessed higher prediction accuracy than conventional prediction models, with an increase in the area under the receiver operating characteristic curve by up to 18.4%. In addition, genome-wide statistical analyses showed that ROSE predictions can be well correlated with diverse putative regulatory factors of ribosome stalling. Moreover, the genome-wide ribosome stalling landscapes of both human and yeast computed by ROSE recovered the functional interplays between ribosome stalling and cotranslational events in protein biogenesis, including protein targeting by the signal recognition particles and protein secondary structure formation. Overall, our study provides a novel method to complement the ribosome profiling techniques and further decipher the complex regulatory mechanisms underlying translation elongation dynamics encoded in the mRNA sequence.

Transcriptome-wide sites of collided ribosomes reveal principles of translational pausing

“…There is little specificity at the E-site, which is notable because previous work that estimated pause sites from monosome footprints identified pausing signatures with a strong E-site bias for proline (e.g. Ingolia et al (2011);Zhang et al (2017); Pop et al (2014). Due to their particular chemistry, prolines (especially in a poly-proline context) are well-known for their difficult peptide 650 bond formation and they are slow decoding, leading to translational stalls that can be resolved through the activity of eIF5A (Gutierrez et al, 2013).…”

“…In combination with quantitative modelling approaches, subsequent studies have identified parameters that can impinge on local translation speed and pausing (reviewed in Schuller and Green (2018)). Among these are, notably, specific amino acids (Charneski and Hurst, 2013), codon pairs (Gamble et al, 2016), tRNA availability (Darnell et al, 2018;Guydosh and Green, 2014), 70 RNA secondary structures (Zhang et al, 2017;Pop et al, 2014), or the folding (Doring et al, 2017) and exit tunnel interactions (Dao Duc and Song, 2018;Charneski and Hurst, 2013) of the nascent peptide. However, to what extent translational pausing occurs in vivo in a mammalian system, which characteristics these pause sites have, and whether they are functionally relevant is still 75 poorly understood.…”

Global ribosome profiling reveals that mutant huntingtin stalls ribosomes and represses protein synthesis independent of fragile X mental retardation protein

“…To identify the mRNA transcripts on which ribosomes are stalled in HD cells, we employed ribosome profiling, a high-throughput sequencing tool that measures ribosome occupancy by sequencing ribosome-protected mRNA fragments (RPF) at a global translatome level (29)(30)(31)(32). Figure 1J shows the research design, where three replicates of wild-type control striatal cells and HD-het and HD-homo mutant striatal cells were subjected to ribosome profiling in which we prepared ribosome footprints (Ribo-Seq) and matching RNA (RNA-Seq).…”

DeepHINT: Understanding HIV-1 integration via deep learning with attention

“…The recent decade has witnessed the boom of deep learning. In computational biology, deep learning has become the state-of-the-art prediction methods in many applications, e.g., identification of nucleotide-protein binding and nucleotide modification sites [16][17][18][19], prediction of the functional effects of noncoding sequence variants [20,21], cancer genomics [22], translation initiation and elongation modeling [23,24] and drug discovery [25,26]. On the other hand, despite the superior prediction performance, the explainability and the understanding of feature organizations of deep learning models often lag behind, which not only limits the applicability of deep learning techniques in exploring unknown cellular mechanisms and gaining insights, but also raises potential concerns of using a black box.…”