COOS-7 contains 132,209 single-cell images of mouse cells, where the task is to predict protein subcellular localization. Images are spread over 1 training set and 4 testing sets, where each single-cell image contains a protein and nucleus fluorescent channels. COOS-7 provides a classification setting where four test datasets have increasing degrees of covariate shift: some images are random subsets of the training data, while others are from experiments reproduced months later and imaged by different instruments. While most classifiers perform well on test datasets similar to the training dataset, all classifiers failed to generalize their performance to datasets with greater covariate shifts. Read more at https://www.alexluresearch.com/publication/coos/.