Supercomputer simulations provide insights into mechanics of deafness

Share this on social media:

A team from Harvard University has resolved the atomic-level structure of a protein essential for sound perception and used TeraGrid supercomputer simulations to determine how it functions in hearing and deafness. The research was published in the journal, Neuron.

The team, led by Rachelle Gaudet and David Corey, used X-ray crystallography to determine the molecular structure of cadherin-23's tip (the protein involved in converting sound into electrical signals), with and without a mutation causing deafness. The team then tested cadherin-23's elasticity using molecular dynamics simulations on TeraGrid systems at the National Center for Supercomputing Applications (NCSA) and the Texas Advanced Computing Center (TACC).

The protein structures were modelled in water boxes to mimic their biological environment; the simulation systems encompassed up to 355,000 atoms. Using the NAMD software developed by the Theoretical and Computational Biophysics Group at the University of Illinois at Urbana-Champaign, the team performed hundreds of molecular dynamics simulations using the Ranger system at TACC and the Abe cluster at NCSA. The powerful machines and large number of processors available allowed them to test the elasticity of cadherin-23's tip in multiple near-physiological conditions. Analysis of more than 5 terabytes of data generated by the simulations revealed an essential role for calcium ions in the mechanics of cadherin-23.

The simulations performed at TeraGrid facilities allowed the team to get insights about the role of calcium ions and cadherin-23 mechanics in hearing and deafness that cannot be obtained with any other technique, and provided predictions that can be experimentally tested.

'The outstanding resources and support provided by the TeraGrid were essential to our work,' said Marcos Sotomayor, a member of the research team at Harvard Medical School. 'We were particularly happy with Ranger and Abe, as we were able to use NAMD on both machines extensively and smoothly. We also benefited from NCSA's Unitree storage system, which was essential for us as we were able to recover all our data when one of our local disks failed.'