The growing scale of HPC is allowing simulation of more and more complex systems at greater detail than ever before, particularly in the biological research spheres. Now, researchers at the University of Stuttgart are leveraging the supercomputer power at the High-Performance Computing Center Stuttgart (HLRS) to conduct a joint analysis of how human bones, musculature and nerves interact to enable our movements.
“Understanding the musculoskeletal system in detail is unfortunately a field characterized by poor access to important information, and making measurements can corrupt datasets, lead to bodily harm, or be too imprecise,” explained Aaron Krämer, one of the University of Stuttgart researchers working on the project, in an interview with HLRS’ Eric Gedenk. “An alternative way of gaining insight in this field, which is becoming common, is to simulate the process of interest. In our case, we are investigating the full activation process from the nervous system to muscle contraction.”
The researchers are working on simulating individual small pieces of these systems, with the aim of eventually scaling up the simulations. For instance, their current target – the human bicep – requires simulation of the dozens of fascicles that constitute the muscle, each in turn containing tens or hundreds of thousands of muscle fibers. Quickly, a “single muscle” turns into billions of calculations.
To run these intensive simulations, the researchers are using the Hawk supercomputer at HLRS. Hawk, an HPE Apollo 9000 system, was inaugurated in February 2020 and includes 5,632 nodes, each equipped with two AMD Epyc Rome 7742 CPUs, and a total of 1.44 petabytes of memory. Hawk delivers 19.3 Linpack petaflops of computing power, placing it 16th on the most recent Top500 list. The team is using around 7,000 cores on Hawk for their simulations.
“Our model is such a multi-scale and multi-physics problem that in order to resolve all the processes – from the subcellular level to what we see when the human body is in motion – supercomputing is essential,” said Miriam Mehl, a professor at the University of Stuttgart and another member of the research team. “We really want to use these highly detailed models, because phenomenological models based on smaller amounts of input data do not give us the same degree of insight or the ability to generalize the observations we are able to see in our models.”
The team believes that models like these may eventually present a pathway for improving drug and treatment development compared to strictly experiment-driven work. To that end, another member of the team – Stuttgart professor Oliver Röhrle – is working with a “Cluster of Excellence” at the university called SimTech, which is working to integrate experimental data with computational modeling.
“The challenge with integrating experimental data and computational models is two-fold. On the one hand, if you have a model informed by a lot of experimental data, you don’t want to over-fit the model to just one data set. At the other end of the spectrum, when modelling something where you don’t have a lot of data, you have to have some kind of model because data can’t tell you the entire story.”
Next, the team is working to incorporate more realistic tendons and electromyography signals (a measure of the electrical activity of muscles) into its simulations, and will soon begin to coupling its individual simulations.
To read the coverage from HLRS’ Eric Gedenk, click here.