Skip to main content
Premium Trial:

Request an Annual Quote

SIB, Quartz Bio to Improve MegaClust's Flow Cytometry Analysis Capabilities for Pharma


Quartz Bio and the Swiss Institute of Bioinformatics are working together to develop and validate MegaClust, SIB's platform for analyzing flow cytomery data, so that they can use it to provide analysis services to pharmaceutical companies.

The partners want to test whether or not MegaClust, which is comprised of a set of clustering algorithms, can be used to "systematically" assess large quantities of data generated from flow cytometry experiments in clinical trials, Jérôme Wojcik, Quartz Bio's CEO, told BioInform.

Quartz spun out of Merck Serono — the biopharmaceutical division of Merck KGaA — last fall, as part of a restructuring at Merck Serono that included the closure of its Geneva headquarters (BI 9/7/2012). Wojcik was formerly director of bioinformatics at Merck Serono, and Quartz was founded to offer a range of services of interest to pharmaceutical firms, with a particular focus on biomarker discovery using data from microarrays, flow cytometry, and other -omics technologies.

Under the new collaboration, Quartz and SIB will assess how much demand there is for tools like MegaClust among pharma companies, said Ioannis Xenarios, director of Vital-IT, SIB's high performance computing and storage division.

If and when the partners decide to offer services around MegaClust, pharma clients can use the tool to stratify patients for clinical trials and also for things like mode of action analysis, Xenarios said.

These firms, Wojcik explained, are increasingly "relying on flow cytometry biomarkers in their stratified medicine approaches," but they aren't able to derive maximum benefits from the technique because of a "lack of unbiased analysis standards [that] impairs the value and the robustness of generated hypotheses."

These experiments, Wojcik explained, can generate thousands of files that could contain a number of mistakes such as sampling errors. Usually, researchers manually check the quality of a subset of those files to ensure that the data they've generated is good, but there might be problems with data in files that they don't check, he said.

That’s where MegaClust comes in. SIB developed it to analyze flow cytometry data in an unbiased fashion, according to Xenarios.

Vital-IT developed MegaClust with Geneva Bioinformatics, a Swiss firm that develops software and databases for screening proteins and small molecules and serves as SIB's technology transfer office.

Vital-IT began working on it about two and half years ago with the help of an innovation grant from the Swiss Commission for Technology and Innovation, Xenarios told BioInform. They spent that time conducting research, developing the algorithm, applying it to datasets internally and in challenges such as Flow Cytometry: Critical Assessment of Population Identification Methods, or FlowCAP, he said.

MegaClust uses unsupervised clustering algorithms to identify and group together cell populations in flow cytometry data that have similar phenotypes or markers, he explained. That’s an improvement over manual analysis approaches, which are difficult for others to reproduce, he said.

Researchers usually analyze data manually when they know what results they want to find in their data, Xenarios explained. For example, a scientist might want to know what fraction of cells in a patient sample has a particular set of markers. Because MegaClust focuses on grouping any sets of cells that are similar, it could discover new previously unknown clusters that the researcher would otherwise have missed, he said.

So far, SIB has tested MegaClust on publicly available datasets but these don't have all of the data quality problems that Quartz Bio has seen in the clinical setting, Wojcik said.

So together, Quartz Bio and SIB will evaluate MegaClust's quality assessment capabilities for about six months using real clinical trial datasets and check its ability to capture different kinds of errors, he said. By partnering, they'll benefit from each others' expertise in clustering methods and clinical flow cytometry data analysis instead of having to become expert in both areas, he said.

SIB will also benefit from Quartz's access to pharma companies who might be interested in taking advantage of MegaClust when they bring the service to market, he said.

If their tests are successful, the partners could sign a second agreement that will allow them to start co-offering services to pharma customers as early as this summer, Wojcik said.

The partners haven’t determined a pricing scheme yet, but they said that it will depend on the number of samples and the complexity of the project.

Quartz's partnership with SIB is the first of other planned collaborations between Quartz and other academic groups that are not being disclosed at this time, Wojcik said.

The company is also hiring, according to Wojcik. Specifically, Quartz Bio is looking to hire a statistics director in April and a bioinformatics analyst in June, he said.