Table 1.

Computational performance of SR and NP

Network			Load time		SR (10 000 r)		NP (10 000 p*)
Type	Genes	Edges	D	HS	D	HS	D	HS**
PPI	12 648	167 445	1.6 s	1.6 s	22 s	33 s	∼2.2 h	∼0.5 h
CO	14 449	8 482 112	59.3 s	50.5 s	29 s	38 s	∼9 d	∼45 h

For two networks [protein–protein interaction (PPI); Co-expression (CO)], we timed SR and NP with 50 seed-genes, for 10 000 randomizations (r)/permutations (p), on a desktop (D) and high-specification (HS) computer. D: 3.6-GB RAM and two 3.16-GHz Intel Core2 Duo CPUs. HS: 148-GB RAM and 24 2.67-GHz Intel Xeon CPUs.*Ten switches per permutation. **Timed using ‘parallelize’ option (see user manual).