Evolutionary sequence modeling for discovery of peptide hormones.

Area of Science:

• Genomics
• Computational Biology
• Neuroscience

Background:

• Many G-protein-coupled receptors (GPCRs) are 'orphan' due to unknown endogenous peptide hormone ligands.
• Identifying these peptide hormones is crucial for understanding cellular signaling and developing new therapeutics.

Purpose of the Study:

• To present a novel computational framework for discovering unknown peptide hormones.
• To identify new functional molecules, particularly peptide hormones, using cross-genomic sequence comparisons.
• To validate the computational methodology through experimental characterization of a novel neuropeptide.

Main Methods:

• Developed a computational framework integrating genomic spatial structure and evolutionary path structure across species.
• Utilized a hierarchical grammar of evolutionary probabilistic models incorporating structural and evolutionary constraints.
• Performed cross-genomic sequence comparisons and functional-element level alignments of human and non-human proteins.

Main Results:

• The computational method identified 45 out of 54 known prohormones with 44 false positives in a proof-of-concept study.
• Identified a novel putative prohormone with at least four potential neuropeptides by comparing human and mouse proteins.
• Experimentally validated the novel neuropeptide, confirming its presence in specific brain regions and other organs.

Conclusions:

• The species comparison and Hidden Markov Model (HMM)-based computational approach successfully identified a previously undiscovered neuropeptide.
• This discovery has significant implications for understanding GPCR pathways and identifying new drug development targets.
• The computational framework offers a powerful tool for discovering novel functional molecules from whole genome sequences.