Clean / Noisy Data

Noisy data: training set
- sequences may contain errors
- sequences may have been assigned to the wrong family

Clean data: the training set is assumed to be “correct”