Table 7. Error analysis: average feature similarity for error cases on Naïve Bayes.

	Caenorhabditis		Danio rerio		Drosophila		Escherichia coli		Zea mays
Feature	FP	FN	FP	FN	FP	FN	FP	FN	FP	FN
#Instances	1644	72	2167	39	13879	4844	161	9	390	66
Description	0.322	0.320	0.293	0.372	0.250	0.515	0.147	0.172	0.216	0.428
Literature	0.115	0.027	0.440	0.243	0.031	0.471	0.003	0.000	0.013	0.232
Length	0.191	0.567	0.165	0.659	0.143	0.704	0.151	0.556	0.207	0.720
Identity	0.936	0.902	0.954	0.902	0.974	0.854	0.983	0.924	0.962	0.866
AP	0.015	0.018	0.008	0.032	0.027	0.060	0.037	0.167	0.054	0.277
Expect_Value	0.012	0.109	0.019	0.031	0.168	0.365	0.037	0.020	0.055	0.001
CDS_Identity	0.881	0.882	0.924	0.888	0.893	0.852	0.906	0.921	0.868	0.840
CDS_AP	0.018	0.022	0.006	0.032	0.020	0.072	0.022	0.146	0.009	0.413
CDS_Expect	0.458	0.348	0.596	0.299	1.126	0.36	0.753	0.589	0.614	0.056
TRS_Identity	0.403	0.512	0.392	0.345	0.426	0.424	0.430	0.548	0.540	0.840
TRS_AP	0.020	0.042	0.020	0.408	0.032	0.130	0.030	0.262	0.027	0.463
TRS_Expect	2.456	1.312	1.630	0.408	2.061	1.404	1.799	0.144	3.227	0.257

#Instances: number of instances; FP: false positives, distinct pairs classified as duplicates; FN: false negatives, duplicates classified as distinct pairs; Feature names are explained in Table 3; Numbers are averages, excluding pairs not have specific features.