Knowledge-based analysis of proteomics data

BMC Bioinformatics

Table 1 Interactome overconnectivity analysis of serum in ovarian cancer.

Object name	Actual	n	R	N	Expected	Ratio	p-value	z-score
Transcription factors
Nkx5-1	2	79	9	19603	0.036	55.14	5.668E-04	10.33
NFIC	4	79	104	19603	0.420	9.54	8.277E-04	5.56
USF2	7	79	183	19603	0.738	9.49	9.061E-06	7.34
Receptors
ITGB6	2	79	9	19603	0.036	55.14	5.668E-04	10.33
VLDLR	3	79	22	19603	0.089	33.84	9.180E-05	9.8
Glypican-1	2	79	16	19603	0.064	31.02	1.855E-03	7.64
Ligands
LAMA2	2	79	7	19603	0.028	70.9	3.324E-04	11.77
CCL17	2	79	11	19603	0.044	45.12	8.614E-04	9.31
Thrombospondin 2	2	79	15	19603	0.060	33.09	1.627E-03	7.91
Kinases
LRRK1	2	79	5	19603	0.020	99.26	1.591E-04	13.98
Proteases
CTRL	1	79	1	19603	0.004	248.1	4.030E-03	15.72
Mcpt8	1	79	1	19603	0.004	248.1	4.030E-03	15.72
CRIM2	1	79	2	19603	0.008	124.1	8.044E-03	11.07

Actual = number of network objects in the activated dataset(s) which interact with the chosen object listed in object name column; n= number of network objects in the activated dataset( i.e. proteomics list); R=number of network objects in the complete database or background list which interact with the chosen object; N= total number of gene-based objects in the complete database or background list; Expected= mean of hypergeometric distribution (n·R/N); Ratio= connectivity ratio (Actual/Expected); z-score= (Actual-Expected)/s.d.(Expected), where s.d. - standard deviation of hypergeometric distribution, p-value= probability to have the given or higher (lower for negative z-score) value of Actual by chance under null hypothesis of no over- or under-connectivity.

ISSN: 1471-2105