OCTOPUS

OCTOPUS predicts transmembrane protein topology and segment boundaries to identify membrane-spanning helices, reentrant/membrane-dipping regions, and transmembrane hairpins for structural and functional analysis of proteins that constitute approximately 25% of a typical genome.

Key Features:

• Hybrid modeling: Combines hidden Markov models (HMMs) and artificial neural networks (ANNs) in its prediction engine.
• Complex topology handling: Integrates reentrant/membrane-dipping regions and transmembrane hairpins into topology predictions.
• Topology and segment prediction: Delivers residue-level identification of membrane-spanning helices and topology states.
• Benchmark performance: Demonstrated a 94% accuracy rate on a benchmark of 124 sequences with known structures.

Scientific Applications:

• Structural analysis of membrane proteins: Provides topology maps to support interpretation of membrane protein structure and folding.
• Functional characterization and target evaluation: Facilitates investigation of transmembrane protein roles in cellular processes and aids evaluation of therapeutic targets.

Methodology:

Uses a hybrid computational approach combining hidden Markov models (HMMs) and artificial neural networks (ANNs); a benchmark on 124 sequences with known structures reported 94% accuracy.