TABLE VIII.

The ranks and the p-values of the target pathways for SPIA (all genes) and pDis analysis (all genes) using 8 yeast knock-out gene data sets. The improvement factor based on Fig. 2 is calculated for each data set considering 5% significance threshold using FDR-corrected p-values. The average improvement factor shows that pDis analysis (all genes) improves the results 50% compared to SPIA (all genes). Three target pathways were found to be significant after FDR-correction in pDis analysis (all genes) while only one target pathways have significant FDR-corrected p-values in SPIA (all genes). As shown, the average p-value and rank for the target pathways are lower (i.e. better) in the pDis analysis (all genes) than in SPIA (all genes). The results show that pDis analysis (all genes) yields significantly better p-values than SPIA (all genes) for the target pathways (p-value from t.test = 0.002) as well as it has significantly lower ranks for the target pathways compared to SPIA (all genes) (p-value from Wilcoxon test = 0.01).

	knock-out gens	target pathway	SPIA (all genes)			pDis analysis (all genes)			improvement factors compared to SPIA
			p-values	FDR	ranks	p-values	FDR	ranks
1	APC9	Cell cycle Meiosis - yeast	0.422 0.917	1 1	3 5	0.133 0.184	0.411 0.414	2 4	+0.5 +0.5
2	TPK3	Meiosis- yeast	5e-06	4.5e-05	1	5e-06	4e-05	1	–
3	RGT2	Meiosis- yeast	0.075	0.225	3	0.001	0.009	1	+1
4	USA1	Protein processing in endoplasmic reticulum	1	1	7.5	0.092	0.27	3	+0.5
5	TIF4631	RNA transport	1	1	7.5	0.084	0.756	1	–
6	URM1	Sulfur relay system	0.048	0.306	1	0.040	0.36	1	+1
7	SSM4	Protein processing in endoplasmic reticulum	1	1	7.5	0.004	0.018	2	+0.5
8	CUE1	Protein processing in endoplasmic reticulum	1	1	7.5	0.208	0.624	3	+0.5
Average			0.606	0.725	4.77	0.082	0.318	2	4.5/9=50%