HHomp--prediction and classification of outer membrane proteins.

Area of Science:

• Structural biology
• Bioinformatics
• Genomics

Background:

• Outer membrane proteins (OMPs) are crucial transmembrane proteins in Gram-negative bacteria, mitochondria, and plastids.
• Existing prediction methods often rely on analogous features like hydrophobicity patterns.
• OMPs are frequently related through common ancestry, a feature not fully exploited by current tools.

Purpose of the Study:

• To develop a novel computational method for accurate OMP identification.
• To utilize conserved homologous relationships among OMPs for improved prediction.
• To create a comprehensive database of OMP profile Hidden Markov Models (HMMs).

Main Methods:

• HHomp utilizes profile HMMs and pairwise HMM comparisons against an extensive OMP database.
• The OMP database was built using HHsenser, a homology detection method, starting from known OMPs.
• Predictions are integrated with PROFtmb for enhanced accuracy.

Main Results:

• HHomp achieves a 63.5% true positive rate before the first false positive on the TransportDB benchmark.
• This represents a significant improvement over existing methods like PROFtmb, BOMP, and TMB-Hunt.
• HHomp accurately identified 57 out of 59 known OMPs in Escherichia coli, classifying them into functional subgroups.

Conclusions:

• HHomp offers a highly accurate and sensitive method for identifying outer membrane proteins based on homology.
• The approach significantly outperforms current state-of-the-art OMP prediction tools.
• This method aids in understanding bacterial outer membrane protein functions and evolution.