Unveiling Contacts within Macromolecular Assemblies by Solving Minimum Weight Connectivity Inference (MWC) Problems.

Area of Science:

• Biochemistry
• Structural Biology
• Computational Biology

Background:

• Macromolecular assemblies are formed by interacting protein subunits (oligomers).
• Inferring contacts between these subunits is crucial for understanding their function.
• Existing methods for connectivity inference (CI) have limitations in accuracy and incorporating prior knowledge.

Purpose of the Study:

• To develop novel algorithms for minimum weight connectivity inference (MWCI) and bootstrap-enhanced connectivity inference.
• To improve the sensitivity and specificity of predicting subunit contacts in macromolecular assemblies.
• To provide a robust computational tool for analyzing oligomer data.

Main Methods:

• Developed two algorithms: MILP-W for minimum weight connectivity inference and MILP-WB using a bootstrap approach.
• Applied algorithms to analyze oligomer data from yeast exosome, yeast proteasome lid, and human eIF3 complexes.
• Validated predictions against known contact information from crystal structures, cryo-electron microscopy, and cross-linking data.

Main Results:

• The MILP-W and MILP-WB algorithms demonstrated high specificity and sensitivity in predicting subunit contacts.
• Achieved a significant improvement over previous methods, with typically a twofold increase in sensitivity.
• Successfully identified known contacts in well-characterized biological systems.

Conclusions:

• The novel MILP-W and MILP-WB algorithms offer a powerful and accurate approach for connectivity inference from oligomer data.
• These methods represent a substantial advancement in the field, enhancing our ability to decipher complex macromolecular structures.
• The accompanying software is expected to be widely useful for researchers facing similar connectivity inference challenges.