We are developing new models to improve sequence similarity based annotations of proteins. We have created unique models to predict function relatedness and specificity based on sequence similarity. This work is funded through an NSF grant.