FREEDA: An automated computational pipeline guides experimental testing of protein innovation

Cell biologists typically focus on conserved regions of a protein, overlooking innovations that can shape its function over evolutionary time. Computational analyses can reveal potential innovations by detecting statistical signatures of positive selection that lead to rapid accumulation of beneficial mutations. However, these approaches are not easily accessible to non-specialists, limiting their use in cell biology. Here, we present an automated computational pipeline FREEDA that provides a simple graphical user interface requiring only a gene name; integrates widely used molecular evolution tools to detect positive selection in rodents, primates, carnivores, birds, and flies; and maps results onto protein structures predicted by AlphaFold. Applying FREEDA to >100 centromere proteins, we find statistical evidence of positive selection within loops and turns of ancient domains, suggesting innovation of essential functions. As a proof-of-principle experiment, we show innovation in centromere binding of mouse CENP-O. Overall, we provide an accessible computational tool to guide cell biology research and apply it to experimentally demonstrate functional innovation.