Predicting disease associations via biological network analysis

BMC Bioinformatics

Table 2 Evaluation of our measures against ICD-9 classification

Data	Group	Annotation-based	Function-based	Topology-based
OMIM	Same	0.0114 ± 0.0665	0.0355 ± 0.0892	0.4349 ± 0.1101
	Different	0.0010 ± 0.0139	0.0118 ± 0.0314	0.3996 ± 0.0760
	P-value	1.2785 ×10⁻¹³	1.0423 ×10⁻⁵²	2.1257 ×10⁻⁵⁴
CTD	Same	0.0361 ± 0.1590	0.0728 ± 0.1754	0.4863 ± 0.1770
	Different	0.0050 ± 0.0274	0.0333 ± 0.0662	0.4408 ± 0.1368
	P-value	1.4887 ×10⁻²³	1.4040 ×10⁻⁹	2.0240 ×10⁻²⁵
FunDO	Same	0.0418 ± 0.1344	0.0991 ± 0.1611	0.5560 ± 0.2214
	Different	0.0100 ± 0.0262	0.0549 ± 0.0830	0.4952 ± 0.1636
	P-value	1.7609 ×10⁻¹⁴⁴	9.6708 ×10⁻¹⁰⁰	2.7037 ×10⁻⁹⁰
HuGENet	Same	0.0931 ± 0.1798	0.2470 ± 0.2123	0.8031 ± 0.2248
	Different	0.0438 ± 0.0566	0.1881 ± 0.1522	0.7837 ± 0.2292
	P-value	1.4585 ×10⁻⁷⁴	9.9053 ×10⁻⁷²	4.5910 ×10⁻¹⁴
Intersection	Same	0.0338 ± 0.1511	0.0593 ± 0.1907	0.3826 ± 0.1131
	Different	0.0024 ± 0.0329	0.0089 ± 0.0428	0.3496 ± 0.1020
	P-value	2.2667 ×10⁻²	2.7448 ×10⁻⁴	5.4716 ×10⁻⁴
Union	Same	0.0350 ± 0.1179	0.0963 ± 0.1463	0.5680 ± 0.2226
	Different	0.0085 ± 0.0219	0.0583 ± 0.0818	0.5042 ± 0.1716
	P-value	1.3493 ×10⁻²¹¹	7.1478 ×10⁻¹¹³	4.1709 ×10⁻¹⁴¹

Numbers in the table are similarity scores between diseases from the same ICD-9 categories, compared with those from different ICD-9 categories. P-values are calculated by using the Mann −Whitney U test.

ISSN: 1471-2105