FuSe

FuSe groups mRNA transcripts by predicted protein functional similarity to enable functional-level interpretation of RNA-Seq data.

Key Features:

• Functional grouping of mRNA transcripts: Groups transcripts based on predicted protein functional similarity instead of solely on chromosomal location, gene loci, or transcript identifiers.
• Protein structure–based prediction: Leverages primary and secondary structures of proteins to predict functional similarity among transcript-encoded proteins.
• Confidence scoring: Computes Knowledge Score (KS) and Discovery Score (DS) for protein pairs to estimate the likelihood of shared function.
• Similar function protein groups: Aggregates protein pairs with high KS/DS confidence into "similar function protein groups" for downstream analysis.
• Expression aggregation at functional level: Performs expression calculations aggregated at the level of similar function protein groups rather than at individual transcripts or genes.
• Analytical reframing: Replaces traditional chromosomal or gene loci-based and transcript-level assumptions with functional similarity–based analysis to address cases where genes encode multiple transcript types with varying protein-coding potential.

Scientific Applications:

• Paracetamol exposure RNA-Seq: Applied to in vitro cells exposed to paracetamol, FuSe identified differentially expressed genes involved in cell adhesion and glycogen regulation that were not detected by conventional analyses.
• Data provenance: Analyses related to the paracetamol studies are associated with BioStudies accession numbers S-HECA143, S-HECA158, and S-HECA139, with supplementary data reported in Bioinformatics.

Methodology:

Predict functional similarity using protein primary and secondary structure, compute KS and DS for protein pairs, group protein pairs exceeding confidence thresholds into similar function protein groups, and aggregate expression at the functional-group level.