Use case: Lisa processes data of millions of mouth bacteria
Most people would fail to associate dentistry with High Performance Computing. However, appearances are deceptive: as of 2006, the technique of sequencing - whereby bacteria are identified on the basis of genetic information - is now widely applied by dentistry researchers.
Dr. Egija Zaura of the preventive dentistry department at the Academic Centre for Dentistry in Amsterdam (ACTA) - an alliance between the University of Amsterdam (UvA) and VU University Amsterdam (VU) - explains the latest developments. ACTA researchers originally used simple in vitro models, but this approach yielded insufficient insight into the complex interactions between oral bacteria. “We started investigating the various available tools”, Zaura explains. “In 2006, 454 pyrosequencing became available - a technique that allows for the parallel sequencing of millions of bacteria. You can then apply this data to identify the various strains of bacteria in a patient's mouth and determine their ratios. We were the first dentistry department in the world to apply this technique to clinical dentistry samples.”
Human Microbiome Project
The technique showed so much promise that ACTA wanted the capacity to conduct its own analyses, Zaura explains: “We contacted SURFsara in 2010 and started our analyses. We apply knowledge developed as part of the Human Microbiome Project (HMP), an initiative aimed at characterizing microbial communities in multiple human body sites. Researchers working on the project develop a large amount of open source material. We use a pipeline (a sequencing tool) developed within the HMP. The pipeline was installed on the Lisa compute cluster at SURFsara, and allows us to conduct graphical analyses.”
“The staff at SURFsara are very accessible and have been very responsive to our needs. The staff have been extremely helpful and creative in finding solutions to the challenges we face.”
ACTA has employed its own in-house bioinformatics expert since 2011. This was no small luxury, in view of the ever-expanding amount of data: “In addition to 16S rRNA gene-based analysis which allows us to identify bacteria, we currently also apply shotgun sequencing”, Zaura explains. “We basically put a sample through the meat grinder and analyze small particles of all the various genes. These particles are then sequenced, resulting in what is known as a metagenome. We can then apply this metagenome to identify the various bacteria's functions. For example, we are currently studying the differences between bacteria from a sample with caries or periodontitis and bacteria from a healthy mouth.”
With the growing amount of data yielded by shotgun sequencing, data processing (and thus computing power) is becoming increasingly important, but storage too is crucial to ACTA's research activities. In addition to the Lisa compute cluster, ACTA also uses SURFsara data storage facilities. Zaura: “The staff at SURFsara are very accessible and have been very responsive to our needs. We also received a great deal of support during the initial phases, while we worked to develop an action plan. The staff have been extremely helpful and creative in finding solutions to the challenges we face.”