Table 2. The impact of prior weight.

Prior weight	Number of prior edges	Selection probability per edge in one run+	Probability of one prior edge remaining in final network by chance++	Expected number of prior edges remaining in final network by chance	Number of prior edges remaining in final network	P-value
1	269	0.537	4.67E-15	1.26E-12	14	9.98E-178*
2	122	0.735	3.85E-05	4.69E-03	4	1.92E-11*
3	67	0.808	9.39E-03	6.29E-01	6	4.19E-05*
4	42	0.834	4.53E-02	1.90	10	1.41E-05*
5	20	0.844	7.53E-02	1.50	4	5.90E-02
6	16	0.848	8.99E-02	1.43	4	4.94E-02*
7	6	0.849	9.59E-02	5.75E-01	0	1
> = 8	53	≈0.850	≈9.81E-02	≈5.20	8	≈1.44E-01

⁺ Selection probability per edge in one run = 0.85(1-e—prior weight).

⁺⁺ Assuming the reliability cutoff is 90, which means one prior network should be appear at least 90 times out of 100 runs. The probability is calculated based on binomial distribution.

* Statistical significance (P-value <0.05).