GUIdEStaR

GUIdEStaR integrates genomic and transcriptomic annotations of G-quadruplexes, upstream open reading frames (uORFs), internal ribosome entry sites (IRESs), RNA and histone methylation, small RNAs, and repetitive sequences to enable consolidated analyses across ~40,000 genes and ~320,000 transcripts.

Key Features:

• Integrated data sources: Aggregates existing databases containing information on G-quadruplexes, uORFs, IRESs, RNA and histone methylation, small RNAs, and repeats.
• Transcript segmentation: Divides each mRNA transcript into five regions: 5' UTR, 3' UTR, exons, introns, and biological regions.
• Presence–absence annotation: Provides presence–absence data for 169 distinct element types within each transcript region.
• Coverage: Covers approximately 40,000 genes and 320,000 transcripts.
• AI-compatible datasets: Generates datasets formatted for neural network and other AI-based analyses, exported in CSV format.
• Example code: Supplies example Java code for working with generated datasets.

Scientific Applications:

• Small RNA-mediated TF classification: Neural network classification of transcription factor genes by mediation from SARS-CoV-2 small RNAs versus human microRNAs (miRNAs).
• Receptor–virus interaction classification: Neural network classification of cell membrane receptor genes by virus interaction status.
• NMD target classification: Neural network classification of cell membrane and nuclear receptor genes as nonsense-mediated mRNA decay (NMD) targets or non-targets.
• Consolidated element analysis: Enables comparative analyses of regulatory elements and epigenetic marks across transcript regions.

Methodology:

Integrates multiple external databases on listed element types; partitions each mRNA into 5' UTR, 3' UTR, exons, introns, and biological regions; computes presence–absence annotations for 169 element types per region; exports datasets in CSV for neural network analyses and provides example Java code.