Menu

Graham School News

CLSSS Features Richard Williams on the Challenges of Transferring Terabytes of Data

On November 15 at UChicago’s Gordon Center for Integrated Sciences, Richard Williams IV, PhD, application developer and computational scientist at the Research Computer Center (RCC), delivered the second lecture in the Computational Life Sciences Seminar Series (CLSSS) in a presentation that explored the archiving, analysis, and publication of X-ray Reconstruction of Moving Morphology (XROMM) datasets using the XROMM Data Management (XDM) system developed at RCC.

A collaboration between the Graham School Master of Science in Biomedical Informatics (BMI) program and the Center for Research Informatics (CRI), the CLSSS has the goal of bringing together chemists, biologists, and clinicians, as well as others who might be siloed across the various life sciences at UChicago, for the presentation of high level research as a way to explore common computational challenges facing workers across these various fields.

“There are a core set of computational challenges everyone faces today working in the biosciences,” said Kyle Hernandez, PhD, research assistant professor and bioinformatician at CRI, as well as CLSSS organizer. “These range across areas such as transferring and storing large datasets to reproducibility and visualization. We were excited to partner with the BMI program as a way to expand the range of experience and backgrounds taking part in the seminars.”

Suzanne Cox, PhD, MP, director of the BMI program, noted that these sorts of collaborations build on a key strength of the Graham School program, which is the access its students have to resources and faculty at the University of Chicago. 

“Many of our students arrive with strong backgrounds in computation eager to expand their bioinformatics and genomics knowledge bases,” she said. “Working together with CRI to put together the seminar series is a great way for biomedical informatics students to learn about the breadth of research and opportunities happening at the University while connecting with others confronting similar challenges.”

Beginning his presentation by outlining the core services offered by RCC, including its hardware, software, and technical support capacity, Williams, who received his PhD in Computational Neuroscience from UChicago in 2015, highlighted the University’s high performance computing cluster, Midway, whose 30,000 cores, 1500 nodes, and 4PB of storage data is available to all UChicago researchers, with nearly 2000 users presently active.

From here, Williams moved on to describe XROMM, a 3D imaging technology developed at Brown University that visualizes rapid skeletal movement at high accuracy with applications in biomechanics, neuromechanics, and orthopedics. The system captures high-frame-rate video from two X-ray sources, thereby allowing its software to triangulate the three-dimensional positions of bones as they move. Williams discussed and shared photographs of UChicago’s XROMM facility, housed in the Carlson Animal Research Lab, and highlighted its mobile X-ray emitters and image intensifiers that enable high customization for particular behavioral experiments.

“XROMM has even been used to capture bird and bat flight,” Williams noted after showing video focused on the jaw kinematics of a masticating pig. “It lets you record behavior while also capturing relevant signals, such as real time neural data. Using high-resolution 3D bone models means motion can be studied in the context of detailed skeletal morphology. In the case of pig mastication, for instance, this means it’s possible to visualize and measure the interactions of specific upper and lower tooth cusps during occlusion.”

The core challenge presented by XROMM, Williams continued, involves transferring the terabytes of data users accumulate while using XROMM to Midway in a safe and secure manner, from where it can then be accessed and searched for further analysis. The centerpiece of his presentation, then, involved a detailed explanation and demo of the XDM console, an application Williams developed at RCC that allows for batch upload of data from XROMM to Midway.

“The XDM system packages trial data and metadata collected at the XROMM facility and transfers it to Midway,” Williams said. “Once it’s there, it can be stored long-term and also accessed using a searchable web interface.”

After a lively question and answer period in which Williams revealed that the XDM system would be online by early 2018 and that RCC is in the process of developing a HIPAA-compliant secure data enclave, audience members were asked to stay tuned for the next CLSSS lecture, which Hernandez pledged would take place in January.