2 posts were split to a new topic: Should I use aligned or non-aliged representative sequences to train my own classifier