Submission

The obesity epidemic is shortening the lifespan of people worldwide. Body weight is affected by differences in energy balance: the ratio of energy intake to energy expenditure. We measure these values with indirect calorimeters to monitor O₂ and CO₂ gas exchange rates of animals. The critical barrier is the large amount of data in different formats which requires specific data processing and analysis for each experiment. Surprisingly, no resources exist to fill these needs. We developed a free online tool, CalR, to help scientists quickly and efficiently analyze calorimetry data and to provide standardized methods for reproducible research. We help scientists to share their data. CalR has been used more than 35,000 times all over the world.

Our team is based out of Harvard Medical School and the Beth Israel Deaconess Medical Center. Alex Banks is an Associate Professor of Medicine and an expert in obesity research. He recognized the data problem facing obesity researchers and recruited help in statistics and programming from William and Pruthvi. Pruthvi is an expert at developing interactive programs for data analysis. William has experience with bioinformatics and the design of user interfaces. Together the CalR Conquerors have developed a program to help with data sharing and data re-use. 

The kernel for the concept of a tool to help analyze the giant datasets produced by indirect calorimeters was first germinated during Dr. Banks’ postdoctoral fellowship. During this time, Dr. Banks had to manually curate thousands of rows of spreadsheet data. It became glaringly obvious that there needed to be a software tool to do this, and that no such tool existed. We began designing CalR in 2015 . We had a series of user group meetings, feedback sessions, and surveys during the development process to better understand the needs of the community. We launched CalR from beta into live use in 2017. The goal has always been to simplify and standardize data processing, to allow for data sharing and data re-use. Since its launch in 2017, CalR has been used more than 35000 times. 

We have implemented a meta-data file to collect the information necessary to interpret the raw data. The standardization of meta-data has been equally important as standardizing the data format. We adopted a system where the original experimental data files are maintained in their original format with no modifications. Any necessary changes or data quality improvements are stored within the metadata file and are readily shared. This two-file system has provided considerable flexibility for our users. For example, if two different cohorts of animals are being examined within the same experimental timeframe. The two groups can be distinguished by the different metadata. 

Our specific use case is notable for changing the practice for an entire entrenched field of research. Indirect calorimetry experiments first became common in the 1990’s. However, no data-sharing policies, systems, or agreements ever arose spontaneously. As relative outsiders, we provided a tool that was fast, and easy and saved everyone hours of manual formatting and plotting. As a relatively underappreciated footnote, the CalR program also provided a standardized data format that allowed for data sharing and re-use. 

CalR solves the long-standing data-sharing problem in obesity and metabolic research. We used three approaches. 

1. Our website is free. Users can upload data from any format and will have IT made into a standardized file format. File uploading, data conversion, and reading are free for anyone, always. No tool like that exists for this domain.
2. Neutrality and consensus building. There are three main manufacturers of indirect calorimetry systems. We have worked with all manufacturers to read files from any of these systems. We have a commitment to work together if file formats evolve over time to maintain compatibility. Their users benefit from our tool and we benefit from data sharing and data re-use. We have not accepted financial support from any of the manufacturers to maintain neutrality and the appearance of impartiality. 
3. Open data format. The open data format for CalR files is just a simple comma-separated value file (CSV). Anyone can open the file and see it with any spreadsheet program. While database files, compressed files, or other files might be smaller, these files can be directly interpreted by users.
4. The website uses a graphical user interface that requires no programming. A point-and-click approach makes it easy for users to unravel the different components of body weight regulation. 

This approach could be used in other domains of science. We’ve identified a data sharing bottleneck, designed, hosted, and launched a free tool that helps users. People love it.

This project has a high potential for community engagement; everybody knows about body weight regulation. From a young age, people learn that what they eat and how much exercise will affect their waistlines. Our project allows users to browse and visualize the changes in body weight caused by experimental interventions in laboratory animals. We envision the public will learn from seeing the effects of different diets, sleep cycles, intermittent fasting/time-restricted feeding, temperatures, or pharmacological agents on food intake, energy expenditure and body weight. Our tool, CalR comes with a graphical user interface and no programming experience is needed.