The testing dataset comprises all the 340 GPI-anchored proteins experimentally annotated in SwissProt and contained in All-GPI-Set Abbreviations: TP = True Positives; Cov = Coverage.
* See legend to Table 1.
** The Harvard server of MemType-2L gave an answer only for 309 out of the 340 positive examples comprised in All-GPI-Set. 262 sequences are correctly predicted in this set. The limits to the indexes scoring the performance of this server are computed in the best case, namely by considering all the non predicted proteins as correctly predicted.