I've met the author of the study, Daniel Himmelstein, who is quite passionate about making information free. Projects in his github account (https://github.com/dhimmel) tend to use a CC0 license. Some of his work involves aggregation of data (e.g., https://github.com/dhimmel/hetionet) that is encumbered and he has put a lot of effort into making it as free as possible. His project carefully documents the license for each data point and he took the time to ask copyright holders that do not provide an explicit license to do so.