because the accuracy of the prediction improves if not only the information of the target sequence is used, but also homologous sequences are included.