Ranked retrieval of segmented nuclei for objective assessment of cancer gene repositioning

BMC Bioinformatics

Table 3 Comparison of the automated method with manual analysis shows acceptable agreement

Dataset	Type	# Nuclei	Manual FRA2	Auto FRA2	Manual HES5	Auto HES5	Agree FRA2	Agree HES5
D1	NCBD	63	0.1132	0.1918	0.1363	0.1321	✓	✓
D2	Cancer	164	0.2689	0.4360	0.0000	0.0000	✓	✓
D3	Cancer	115	0.0050	0.0001	0.0003	0.0000	✓	✓
D4	Cancer	133	0.0000	0.0001	0.0000	0.0000	✓	✓
D5	Cancer	188	0.0237	0.0000	0.0000	0.0000	X	✓
D6	Cancer	147	0.0004	0.0182	0.5239	0.5019	X	✓
D7	Cancer	63	0.0000	0.0000	0.0724	0.0013	✓	X
D8	Cancer	174	0.0002	0.0000	0.3017	0.3576	✓	✓
D9	Cancer	120	0.0000	0.0000	0.0000	0.0000	✓	✓
D10	Cancer	117	0.0001	0.0005	0.0157	0.0201	✓	✓
D11	Cancer	87	0.0001	0.0007	0.0000	0.0286	✓	X
D12	Cancer	153	0.0009	0.0048	0.0000	0.0130	✓	X
D13	Cancer	37	0.0000	0.0004	0.0526	0.9827	✓	✓
						Total:	11/13	10/13
							84.60%	76.90%

^a13 test datasets with their tissue type (NCBD = non-cancerous breast disease) and the number of nuclei screened with a probability above 0.1.

ISSN: 1471-2105