QUADRatlas: the RNA G-quadruplex and RG4-binding proteins database

Abstract

RNA G-quadruplexes (RG4s) are non-canonical, disease-associated post-transcriptional regulators of gene expression whose functions are driven by RNA-binding proteins (RBPs). Being able to explore transcriptome-wide RG4 formation and interaction with RBPs is thus paramount to understanding how they are regulated and exploiting them as potential therapeutic targets. Towards this goal, we present QUADRatlas (https://rg4db.cibio.unitn.it), a database of experimentally-derived and computationally predicted RG4s in the human transcriptome, enriched with biological function and disease associations. As RBPs are key to their function, we mined known interactions of RG4s with such proteins, complemented with an extensive RBP binding sites dataset. Users can thus intersect RG4s with their potential regulators and effectors, enabling the formulation of novel hypotheses on RG4 regulation, function and pathogenicity. To support this capability, we provide analysis tools for predicting whether an RBP can bind RG4s, RG4 enrichment in a gene set, and de novo RG4 prediction. Genome-browser and table views allow exploring, filtering, and downloading the data quickly for individual genes and in batch. QUADRatlas is a significant step forward in our ability to understand the biology of RG4s, offering unmatched data content and enabling the integrated analysis of RG4s and their interactions with RBPs.

Figure 1.

Data and analysis features offered by QUADRatlas. The core of the database consists of experimentally derived and predicted RG4s coupled with an extensive catalog of RG4-binding proteins. These proteins can bind to folded or unfolded RG4s and were obtained by collecting and integrating eight screens for RG4 interactors (see Methods). Functional annotations and eCLIP-derived RBP binding sites are also included to empower hypothesis generation. On top of this rich dataset, QUADRatlas offers several visualizations (e.g. a custom genome browser) and analysis capabilities (e.g. RBP–RG4 overlap analysis), coupled with highly flexible options for downloading the data it contains.

Figure 2.

Prediction of RG4s in the coding human transcriptome. The figure shows the properties of RG4 elements predicted in human mRNAs using the QGRSmapper, pqsfinder, and G4hunter algorithms. (A) counts of RG4s predicted by each tool and the consensus set. (B) mRNA region location of predicted RG4s. (C) normalized score distribution of RG4s predicted by each tool and RG4s in the prediction consensus set. (D) intersection of transcripts (top) and genes (bottom) containing an RG4 in the prediction consensus set and in the RT-stop profiling and rG4-seq datasets.

Figure 3.

QUADRatlas user interface. The figure shows an example search for an RG4-binding protein. (A) Parameters selection box, configured to search for RG4s bound by the query RG4BP within those obtained by high-throughput techniques and predictions. The function will return (B) basic information about the query RG4-BP, (C) diseases related to it, and, if known, its associated biological functions. Then, (D) details about RG4-containing RNAs bound by the RG4-BP are given, and (E) RG4s found in these RNAs are retrieved from the experimental and predicted data in QUADRatlas. All tables can be sorted by the column of choice, filtered by keyword using the filter box, and downloaded by clicking on the dedicated button. Finally, (F) the obtained RG4s can be explored in a genome browser that also presents all the other data contained in QUADRatlas.