How many times have you struggled to interpret messy handwriting or a label on a meal deep in your freezer? It can be a frustrating occurrence.

However, when labeling challenges occur in a laboratory, the consequences can be much more severe. The concern has never been greater with the onset of COVID-19, where misidentified labels could have life-changing outcomes.

A team of researchers within the Department of Biology at Queen’s, including Drs. Robert Colautti, Virginia Walker, Stephen Lougheed and Master’s student Yihan Wu have developed a new, flexible research software program that aims to make sample management more reproducible and less prone to human error.

The program is called baRcodeR. This is how it works: Scientists who work with biological samples might record additional information including date, location, measurements, test results, and other observations. Large collaborative projects, like those tracking COVID-19, can require samples and data to be coordinated among hundreds or even thousands of scientists and students working collaboratively from around the world.

“There are a lot of computational tools in the field of ‘data science’ that allow for reproducible workflows, but these focus on data after it is collected,” says Dr. Colautti, Canada Research Chair in Rapid Evolution. “Our program applies these principles to sample labeling and management. Accurate data collection and sample management are crucial to reliable analysis.”

The development of the software came as the result of three large international research projects by the collaborators.

“All three of us (Drs. Walker, Lougheed and Colautti) were each involved in different, large international collaborative research projects, where data collection and data management were becoming a big issue," says Dr. Colautti. These projects included the Global Garlic Mustard Field Survey, TSFN, a project on sustainable fisheries in Canada’s North, and Bearwatch, a polar bear project. “When discussing these very different projects, we realized there was a common set of problems with sample collection and labeling that we couldn’t address with off-the-shelf software.”

Any error with labeling or data management can have serious consequences. For example, according to Dr. Colautti, a mere one per cent labeling error in the more than 80 million COVID-19 tests conducted worldwide could yield hundreds of thousands of misdiagnoses, including tens of thousands of infected patients erroneously cleared to return to work and regular activity. Human errors at this scale are inevitable, particularly for frontline workers who face the mental challenges that come with working long hours under difficult conditions.

The researchers hope that baRcodeR can help remedy some of these issues and, so far, the free, open-access software has been downloaded over 13,000 times and is already being used south of the border.

“baRcodeR is very much in daily use in our ongoing efforts to conduct COVID-19 research in populations of first responders, frontline health care workers, frontline city workers like bus drivers, and a population of local school children and their families” says Chris Barnes, Director of Clinical and Translational Science Informatics and Technology at the University of Florida.

The article “baRcodeR: An open-source R package for sample labeling” appeared in the June 23 issue of Methods in Ecology and Evolution. The software is available through the Comprehensive R Archine Network (CRAN) and the Colautti Lab website.