Accelerating ab initio phasing with de novo models.

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Ab initio phasing remains a significant challenge in protein crystallography.
• Advancements in computational structure prediction allow for de novo models accurate enough for ab initio phasing.
• Current methods require substantial computational resources, limiting their application.

Purpose of the Study:

• To develop a computationally efficient approach for ab initio phasing using de novo models.
• To reduce the extensive CPU time associated with traditional ab initio phasing methods.
• To accelerate the process of solving the protein phase problem.

Main Methods:

• Integrating molecular replacement during the de novo model generation process.
• Avoiding refinement of suboptimal models and terminating simulations early upon obtaining suitable models.
• Benchmarking the new approach against conventional methods using a dataset of 20 proteins.

Main Results:

• The novel approach demonstrated a speed improvement of over two orders of magnitude compared to conventional methods.
• Molecular replacement solutions were frequently obtained shortly after transitioning from coarse-grained to full-atom models.
• The quality of initial coarse-grained models critically impacts the success of subsequent all-atom refinement and molecular replacement.

Conclusions:

• The optimized method significantly accelerates ab initio phasing with de novo models.
• Early initiation of molecular replacement and selective model refinement are key to computational efficiency.
• Generating high-quality coarse-grained models is essential for successful ab initio phasing via molecular replacement.