Mocca: semi-automatic method for domain hunting

Motivation: Multiple OCCurrences Analysis (Mocca) is a new method for repea t extraction. It is based on the T-Coffee package (Notredame et al., JMB, 3 02, 205-217, 2000). Given a sequence or a set of sequences, and a library o f local alignments, Mocca extracts every segment of sequence homologous to a pre-specified master. The implementation is meant for domain hunting and makes it fast and easy to test for new boundaries or extend known repeats i n an interactive manner. Mocca is designed to deal with highly divergent pr otein repeats (less than 30% amino acid identity) of more than 30 amino aci ds.