NMR data-driven structure determination using NMR-I-TASSER in the CASD-NMR experiment

Area of Science:

• Computational Biology
• Structural Biology
• Biophysics

Background:

• Protein structure determination is crucial for understanding biological function.
• Nuclear Magnetic Resonance (NMR) spectroscopy provides valuable structural data.
• Existing computational methods for protein structure prediction have limitations.

Purpose of the Study:

• To evaluate the performance of NMR-I-TASSER, a novel algorithm combining I-TASSER with NMR data, for protein structure determination.
• To assess the algorithm's accuracy and robustness in the CASD-NMR experiment.
• To compare NMR-I-TASSER with traditional molecular dynamics simulations.

Main Methods:

• NMR-I-TASSER employs a molecular replacement-like strategy, utilizing structural templates from the Protein Data Bank (PDB).
• The algorithm iteratively refines structures using Nuclear Overhauser Effect (NOE) assignments and fragment assembly.
• Multiple templates are used to explore diverse structural topologies.

Main Results:

• Without NOE data, NMR-I-TASSER achieved correct structure topology for 15 out of 20 targets (TM-score > 0.5).
• Incorporating NOE restraints significantly improved model accuracy, with all models achieving a TM-score > 0.5.
• Average Root Mean Square Deviation (RMSD) decreased substantially, from 5.29 Å to 2.14 Å (Round 1) and 3.18 Å to 1.71 Å (Round 2).
• The pipeline demonstrated robustness to NOE assignment errors, performing similarly with raw and refined peak lists.

Conclusions:

• NMR-I-TASSER provides a robust and efficient coarse-grained approach for protein structure determination using NMR data.
• The method complements traditional molecular dynamics simulations by rapidly generating near-native structural frameworks.
• This algorithm facilitates atomic-level structural refinement and advances the field of structural biology.