Suppose that you have a large database of biological strings containing sequences of nucleobases drawn from {A, C, T, G}. Some of these strings contain unusual mutations representing changes in the...

Suppose that you have a large database of biological strings containing sequences of nucleobases drawn from {A, C, T, G}. Some of these strings contain unusual mutations representing changes in the nucleobases. Propose an unsupervised method (i.e., neural architecture) using RNNs in order to detect these mutations.