Tyler O'Neal, Staff Editor MIDDLEWARE

NAU prof Gowanlock wins grant to support training in astroinformatics

A computer scientist focused on astroinformatics Mike Gowanlock, an assistant professor at Northern Arizona University’s School of Informatics, Computing, and Cyber Systems was recently awarded a $411,964 grant from the National Science Foundation (NSF). The award will support Gowanlock’s research in the emerging field of parallel computing architectures needed to process large volumes of data generated by major astronomical surveys, including the Legacy Survey of Space and Time (LSST).

The award will support a new undergraduate course that teaches parallel computing and ensuring that graduates of NAU’s computer science undergraduate program have the skills needed to use future generations of supercomputing systems.

Harnessing the power of parallel computing to process the large data volumes 

Designed to lay the groundwork for the launch of the Vera C. Rubin Observatory on the Cerro Pachón ridge in north-central Chile, the 10-year LSST project will begin in 2024. The groundbreaking survey will transform many areas of astrophysics by delivering huge sets of images that will enable scientists to address some of the most pressing questions about the evolution of the universe and the objects in it.

Computer scientist and astroinformaticist Mike Gowanlock, an assistant professor at Northern Arizona University’s School of Informatics, Computing, and Cyber Systems, was recently awarded a $411,964 grant from the National Science Foundation (NSF), which will support research and educational programs in parallel computing.“Modern scientific instruments are generating enormous volumes of data, and these volumes remain a critical challenge for astronomers,” Gowanlock said. “The LSST will enable us to understand how astronomical objects may change over time. Because many interesting astronomical events are transient in nature, it is critical that data from the LSST are processed quickly to enable follow-up observations of transient phenomena using other telescope facilities. New scalable parallel algorithms are needed to enable astronomers to make sense of large data volumes and carry out time-sensitive research objectives. Consequently, one goal of the project is to create new parallel algorithms that exploit emerging computer architectures to enable scientific discoveries in the era of time-domain astronomy.”

NSF’s most prestigious award in support of junior faculty integrating education and research

The NSF made the award to Gowanlock through its prestigious Faculty Early Career Development (CAREER) Program “in support of early-career faculty who have the potential to serve as academic role models in research and education and to lead advances in the mission of their department or organization. Activities pursued by early-career faculty should build a firm foundation for a lifetime of leadership in integrating education and research.”

The LSST will generate 20 terabytes of data each night. Preprocessed data will be sent downstream to NAU to be analyzed by the Solar System Notification and Alert Processing System (SNAPS), a collaboration between Gowanlock and professor David Trilling in the Department of Astronomy and Planetary Science.

The SNAPS system will actively monitor the solar system to determine whether any interesting astrophysical phenomena are occurring, such as an asteroid that is outgassing or changing color. SNAPS will act as an international clearinghouse for solar system research, where astronomers from around the world will listen to the SNAPS data stream, find potential objects of interest, and follow up on those objects using other telescopes facilities.

Parallel computing is a discipline that examines splitting up large amounts of work to be processed by multiple computer processors at the same time. Recently, special purpose processors, such as graphics processing units (GPUs) have been used to solve general-purpose problems, because they have thousands of processors on a single card and are more power-efficient than standard central processing units (CPUs). The project aims to find new ways to exploit the strengths of differing computer architectures as applied to computer science problems that are integral to SNAPS, including problems in databases, unsupervised machine learning, and outlier detection.

The NSF funding will support Gowanlock’s research at the intersection of computer science and astronomy to enable several LSST science goals and will yield standalone parallel algorithms that can be used by domain scientists in other fields.

Grant to support new undergraduate course, internship program, community outreach

The project will combine research and teaching, integrating several activities to ensure both computer scientists and astronomers receive the necessary training to exploit future generation computer systems.

“GPUs are used in the world’s fastest supercomputers,” Gowanlock said. “The grant will support the development of a new upper-division undergraduate course that teaches parallel computing using GPUs. This course will ensure that graduates of NAU’s computer science undergraduate program have the skills necessary to exploit these new architectures in their future careers and enable recent graduates to apply for jobs that require knowledge of parallel computing and/or background in other systems-related topics.”

In addition, the local community will be engaged through outreach activities that promote science, technology, engineering, and mathematical fields, particularly through activities targeting K-12 students. 

“Computer science is severely lacking in representation from underrepresented groups,” Gowanlock said. Another outcome of the project is the development of internship opportunities that primarily serve underrepresented groups in partnership with Lawrence Livermore National Lab (LLNL). Undergraduate students will first perform parallel and distributed computing research with Gowanlock at NAU and will then intern at LLNL for a summer.

“Internship opportunities enable students to get early job experience and have the potential to lead to full-time employment upon graduation. These opportunities can help retain underrepresented groups in computer science, consequently enabling a more diverse and inclusive workforce.”

HMS computational analysis reveals sources of genetic variations

Tyler O'Neal, Staff Editor MIDDLEWARE

The precise transmission of genetic information from one generation to the next is fundamental to life.

Most of the time, this process unfolds with remarkable accuracy, but when it goes awry, mutations can arise—some of them beneficial, some of them inconsequential, and some of them causing malfunction and disease.

Yet, precisely where and how heritable genetic mutations tend to arise in humans has remained largely unknown.

Now, a new multi-institutional study led by investigators at Harvard Medical School and Brigham and Women’s Hospital has pinpointed nine processes during which most human genetic mutations tend to arise.

The work, published Aug. 12 in Science, is based on an analysis of 400 million rare DNA human variants and represents one of the most comprehensive computational efforts to explore heritable genomic variations.

“Genetic mutations are a rare yet inevitable and, indeed essential, part of the development and propagation of the human species—they create genetic diversity, fuel evolution, and occasionally cause genetic diseases,” said study lead investigator Shamil Sunyaev, professor of biomedical informatics in the Blavatnik Institute at HMS and professor of medicine at Brigham and Women’s.  Tuckraider/iStock / Getty Images Plus

“Harnessing the power of computation and big data, we analyzed genomic variations and identified a set of biologic processes responsible for the vast majority of heritable human mutations,” added Sunyaev, who conducted the work with lead authors Vladimir Seplyarskiy, HMS research fellow in medicine at Brigham and Women’s, and Ruslan Soldatov, instructor in biomedical informatics at HMS. 

Key findings
The research identified new mutation-fueling mechanisms and some that were already known. One mechanism was related to inaccurate copying of DNA, another was related to chemical damage occurring to the DNA. The analysis also pinpointed machinery involved in human gene regulation as a frequent culprit in mutations. This machinery is particularly active during early embryonic development, and most of the mutations introduced by the machinery occur during this period. In one surprising finding, the researchers identified a mutation-driving mechanism unrelated to DNA copying and cellular division—processes prone to mutation-causing glitches. This previously unsuspected mechanism leads to mutations in egg cells stored in the ovaries.  

Relevance and implications
The researchers are now working to incorporate some of the results in a model of human-mutation rate along the genome to help predict the chance that a specific mutation would occur at a specific location in the genome. The goal is to help in the analysis of disease mutations and the discovery of genes causing rare diseases. The model may also serve to highlight genes of key importance to human health and survival.

Utrecht University professors visualize the ‘molecular postman’ that delivers important proteins

Tyler O'Neal, Staff Editor MIDDLEWARE

Researchers from Utrecht University, one of the oldest universities in the Netherlands, have visualized the "molecular postman" that makes sure that many important proteins are delivered to the outside of the cell. This has profound implications for our understanding of protein secretion, which in the future can be used for the production of cheaper protein-based therapeutics such as insulin, and the development of new drugs against Corona-, Dengue-, and Zika-viruses as well as novel antibiotics. The researchers are publishing their results today in the academic journal Molecular Cell.

Many proteins, like antibodies, hemoglobin, and some hormones, are produced on the inside of cells and must subsequently be trafficked to the exterior so they can function properly. Like in a post office, cells attach ‘address labels’ called signal peptides to these protein molecules to send them where they are needed. However, in a ‘cellular post office’, some of these address labels need to be removed for the cargo to function properly, while some others must remain, with equally detrimental consequences in case of failure. Therefore, the distinction between these cases is crucial for the function of all sorts of life-sustaining molecules. The 'molecular postman' in action. A - The SPC has to tell signal peptide-containing proteins from others. B - The SPC uses the phospholipid membrane as a 'molecular ruler' to measure the length of signal peptides (pink).

Molecular postman

Using a Nobel prize-winning, state-of-the-art imaging technique called cryo-electron microscopy (cryo-EM), the team of scientists lead by Prof. Friedrich Förster, in collaboration with Prof. Richard Scheltema and his group, has now visualized the human Signal Peptidase Complex (SPC), the ‘molecular postman’ that is at the center of this decision-making process and has to process more than 3,000 different cargo proteins. The research has revealed that the SPC uses a ‘molecular ruler’ made of phospholipids to measure the length of the signal peptides to decide which address labels to remove and which ones to keep.

Formidable challenge

Lead author Manuel Liaci explains: “Visualising the molecular architecture of the SPC was a formidable challenge that had been unmet for decades. The SPC is extremely small and it is integrated into the membrane of the cell, which makes it incredibly difficult to look at with the sub-nanometer resolution we needed.” Prof. Förster and his team employed a multidisciplinary approach combining their imaging technique with mass spectrometry and molecular dynamics simulations on a supercomputer, which enabled them to precisely define the molecular mechanism of signal peptide recognition and removal. The revelation of the SPC structure represents a major advancement in the field since it is the smallest membrane protein that has been visualized to near-atomic detail by cryo-EM so far.