Predicting a sequence based on a single sample