Table 1.

Comparisons of network characteristics in PPI network between aging genes and other groups of genes

Network characteristics	P-valuea and effect sizeb (in parenthesis) of comparison with aging genes				P-valuec of degree-preserved resampling test
	BG	ES	TF	HK	Mean	Median
Degree	4.80 × 10−44 (0.72)	9.74 × 10−20 (0.65)	5.77 × 10−20 (0.66)	2.81 × 10−32 (0.69)	NA	NA
Closeness	2.84 × 10−32 (0.72)	3.22 × 10−16 (0.66)	1.03 × 10−16 (0.67)	1.05 × 10−23 (0.69)	0.396	0.193
Betweenness	3.90 × 10−38 (0.72)	2.12 × 10−16 (0.65)	1.46 × 10−20 (0.68)	2.90 × 10−27 (0.69)	0.627	0.643
Clustering coefficient	1.58 × 10−14 (0.57)	1.70 × 10−05 (0.53)	2.35 × 10−05 (0.53)	2.03 × 10−15 (0.58)	0.005	0.00
Length of shortest path	0.00 (0.62)	0.00 (0.56)	0.00 (0.54)	0.00 (0.58)	0.00	NA

^aWilcoxon rank sum test.

^bCommon language effect size (12) was used to show the difference between groups. Briefly, the effect size was calculated by the proportion of the pairs of values that support a hypothesis among all possible pairs of values between the two groups.

^cSee Materials and Methods section for resampling tests.

The median length of shortest path is not sensitive enough to detect the differences between aging genes and random genes, so the result is not shown here.

Alternative hypothesis: aging genes have higher ‘degree’, ‘closeness’, ‘betweenness’, ‘clustering coefficient’ and ‘shorter shortest path’ than other group of genes and random genes with similar degree distribution as aging genes.

BG, background genes; ES, essential genes; TF, transcription factor genes; HK, housekeeping genes.