A Novel Biclustering Algorithm for the Discovery of Meaningful Biological Correlations between microRNAs and their Target Genes

Table 3 miRTarBase_fil_t results

α	$N (% / %)$	level 1				max level					best level					time (s)
		#cc	p _MF	p _BP	µ _q	lev	#cc	p _MF	p _BP	µ _q	lev	#cc	p _MF	p _BP	µ _q
HOCCLUS2
0.1	12/36	72	0.000	0.000	1.00	6	5	0.099	0.125	0.14	1	72	0.000	0.000	1.00	9
0.2						5	10	0.011	0.015	0.27	1	72	0.000	0.000	1.00	8
0.3						4	17	0.009	0.002	0.41	1	72	0.000	0.000	1.00	8
0.4						4	24	0.000	0.002	0.55	1	72	0.000	0.000	1.00	8
0.5						3	28	0.000	0.001	0.61	1	72	0.000	0.000	1.00	8
HOCCLUS (with METIS)
0.1	100/100	72	0.000	0.015	0.21	3	43	0.008	0.043	0.10	2	47	0.003	0.008	0.11	30
0.2						2	68	0.000	0.006	0.14	2	68	0.000	0.006	0.14	54
0.3						1	72	0.000	0.015	0.21	1	72	0.000	0.015	0.21	70
0.4						1	72	0.000	0.015	0.21	1	72	0.000	0.015	0.21	71
0.5						1	72	0.000	0.015	0.21	1	72	0.000	0.015	0.21	71
ROCC
-	41/16	198	1.000	1.000	0.01	-	-	-	-	-	-	-	-	-	-	89

miRTarBase_filt results. $N (% / %)$ represents the percentage of biclustered mRNAs and miRNAs. #cc is the number of biclusters. lev represents the level number. Results do not vary at the first level since it does not depend on the α value. β does not influence results because of binary interaction scores.

ISSN: 1471-2105