Abstract:ABSTRACT Background: Short-read RNA sequencing (RNAseq) has widely been used to sequence RNA from a wide range of different tissues, developmental stages and species. However, the technology is limited by inherent biases and its inability to capture full-length transcripts. Long-read RNAseq overcomes these issues by providing reads that can span multiple exons, resolve complex repetitive regions and the capability to cover entire transcripts. Unfortunately, this technology is still prone to higher error rates. Noncoding RNA transcripts are highly specific to different cell types and tissues and remain underrepresented in current reference annotations. This problem is exacerbated by the dismissal of sequenced reads that align to genomic regions that do not contain annotated transcripts, resulting in approximately half of the expressed transcripts being overlooked in transcriptional studies. Results: We have developed a pipeline, named HyDRA (Hybrid de novo RNA assembly), which combines the precision of short reads with the structural resolution of long reads, enhancing the accuracy and reliability of custom transcriptome assemblies. Deep, short- and long-read RNAseq data derived from ovarian and fallopian tube samples were used to develop, validate and assess the efficacy of HyDRA. We identified more than 50,000 high-confidence long noncoding RNAs, most of which have not been previously detected using traditional methods. Conclusions: HyDRA's assembly performed more than 40% better than a similar assembly obtained with the top-ranked stand-alone de novo transcriptome short-read-only assembly tool and over 30% better than one obtained with the best-in-class multistep short-read-only approach. Although long-read sequencing is rapidly advancing, the vast availability of short-read RNAseq data will ensure that hybrid approaches like the one implemented in HyDRA continue to be relevant, allowing the discovery of high-confidence transcripts within specific cell types and tissues. As the practice of performing hybrid de novo transcriptome assemblies becomes commonplace, HyDRA will advance the annotation of coding and noncoding transcripts and expand our knowledge of the noncoding genome.

HydRA: Deep-learning models for predicting RNA-binding capacity from protein interaction association context and protein sequence

Dynamic characterization and interpretation for protein-RNA interactions across diverse cellular conditions using HDRNet

A deep learning framework for modeling structural features of RNA-binding protein targets

MAHyNet: Parallel Hybrid Network for RNA-Protein Binding Sites Prediction Based on Multi-Head Attention and Expectation Pooling

Prediction of RNA-protein Sequence and Structure Binding Preferences Using Deep Convolutional and Recurrent Neural Networks

Predicting Dynamic Cellular Protein–rna Interactions by Deep Learning Using in Vivo RNA Structures

HybridRNAbind: Prediction of RNA Interacting Residues Across Structure-Annotated and Disorder-Annotated Proteins

High-throughput discovery of regulatory effector domains in human RNA-binding proteins

DEEPStack-RBP: Accurate identification of RNA-binding proteins based on autoencoder feature selection and deep stacking ensemble classifier

DeepPN: a deep parallel neural network based on convolutional neural network and graph convolutional network for predicting RNA-protein binding sites

A deep learning framework to predict binding preference of RNA constituents on protein surface

Seq-RBPPred: Predicting RNA-Binding Proteins from Sequence.

Improved prediction of DNA and RNA binding proteins with deep learning models

Predicting circRNA-RBP Binding Sites Using a Hybrid Deep Neural Network

DeepRCI: Predicting RNA-chromatin Interactions Via Deep Learning with Multi-Omics Data

Efficient mapping of RNA‐binding residues in RNA‐binding proteins using local sequence features of binding site residues in protein‐RNA complexes

HyDRA: a pipeline for integrating long- and short-read RNAseq data for custom transcriptome assembly

Predicting dynamic cellular protein-RNA interactions using deep learning and <i>in vivo</i> RNA structure

DRBPPred-GAT: Accurate prediction of DNA-binding proteins and RNA-binding proteins based on graph multi-head attention network

RNA Secondary Structure Representation Network for RNA-proteins Binding Prediction

RBPBind: Quantitative Prediction of Protein-RNA Interactions