QGRS-H Predictor: a web server for predicting homologous quadruplex forming G-rich sequence motifs in nucleotide sequences

Abstract

Naturally occurring G-quadruplex structural motifs, formed by guanine-rich nucleic acids, have been reported in telomeric, promoter and transcribed regions of mammalian genomes. G-quadruplex structures have received significant attention because of growing evidence for their role in important biological processes, human disease and as therapeutic targets. Lately, there has been much interest in the potential roles of RNA G-quadruplexes as cis-regulatory elements of post-transcriptional gene expression. Large-scale computational genomics studies on G-quadruplexes have difficulty validating their predictions without laborious testing in 'wet' labs. We have developed a bioinformatics tool, QGRS-H Predictor that can map and analyze conserved putative Quadruplex forming 'G'-Rich Sequences (QGRS) in mRNAs, ncRNAs and other nucleotide sequences, e.g. promoter, telomeric and gene flanking regions. Identifying conserved regulatory motifs helps validate computations and enhances accuracy of predictions. The QGRS-H Predictor is particularly useful for mapping homologous G-quadruplex forming sequences as cis-regulatory elements in the context of 5'- and 3'-untranslated regions, and CDS sections of aligned mRNA sequences. QGRS-H Predictor features highly interactive graphic representation of the data. It is a unique and user-friendly application that provides many options for defining and studying G-quadruplexes. The QGRS-H Predictor can be freely accessed at: http://quadruplex.ramapo.edu/qgrs/app/start.

Figure 1.

Computational workflow of QGRS-H Predictor. The QGRS Predictor performs a three-stage computation to produce homology results given two sequences. The results of the QGRS identification stage (performed individually on each sequence) is combined with the results of the semi-global alignment stage to perform the last stage, homology computation. The results of the last stage are filtered according to settings specified in the homology map and presented to the user.

Figure 2.

Input fields for entering the pair of nucleotide sequences in a variety of formats. The program accepts four modes of data entry, NCBI accession, GI number, FASTA and the raw format. Using the first two methods, nucleotide sequence data, along with annotations and sequence features, will be automatically downloaded and utilized. When entering FASTA format, the program will always use the sequence data entered, but will still download the associated meta-data using the provided accession number.

Figure 3.

Results page. The top panel can be expanded to show details about the sequences entered, including species, common names and sequence features such as UTRs, CDS and poly(A) sites. The second panel contains a homology map that can be used to filter the homology pairing results (bottom of the page) according to their locations and characteristics. The third and fourth panels provide a sequence viewer and a tabular listing of the homology results.

Figure 4.

The highly interactive homology map allows the user to click and drag to zoom into regions of interest within the aligned sequences. The homology map provides an overview of the regions of the principal sequence containing highly conserved QGRS instances. The vertical bars represent the maximum and average homology score of all QGRS pairs found at the given location within the sequence. The user can highlight individual or contiguous sets of regions by dragging the mouse on them, which will filter the results table accordingly.

Figure 5.

Filtering controls for defining characteristics of homologous QGRS motifs. Homology results can be filtered based on their characteristics. The top slider controls the minimum homology score to be shown (a homology score of 1 is the highest value possible). The other sliders control the minimum number of tetrads and minimum G-score of which each QGRS instance in a pair must exhibit in order to be included in the results. (A minimum default G-score of 13 is necessary to weed out potentially false positive instances of QGRS motifs. Please see ‘Methods’ section for details).