Table 1.

Sensitivity of remote homology identification based on different sequence descriptors

Descriptor class	Sequence descriptor	Sensitivitya
Global	PSI-BLASTb	123/445 = 27.6%
	FASTAc	30/445 = 6.7%
	SSEA	95/445 = 21.3%
	AAC	32/445 = 7.1%
	DPC	17/445 = 3.8%
Nonlocal	ACCT_AA	20/445 = 4.5%
	ACCT_SS	55/445 = 12.4%
	CTD_AA	14/445 = 3.1%
	CTD_SS	33/445 = 7.4%
	Triplet	12/445 = 2.7%
Local	Motif_SCOPd	94/445 = 21.1%
	Motif_CATHe	69/445 = 15.5%
	ISITES_CATHf	17/445 = 3.8%

^a The sensitivity was defined as the percentage of correctly assigned remote homologous protein pairs.

^b The modified e-value from PSI-BLAST searching was taken as the descriptor for the similarity between two evaluated sequences.

^c The opt alignment score from the FASTA alignment was used as the descriptor.

^d The term MOTIF_SCOP_SIM, described in equation 9, was used to measure the similarity.

^e The descriptor was based on MOTIF_CATH_SIM.

^f The descriptor was based on ISITES_CATH_SIM.