Table 5. Hamming distance for percent tumor invasion results.

	SuMO-Fil	Low Means	Low Variance
Unfiltered	M = 192, R = 20	M = 1293, R = 64	M = 1831, R = 47
SuMO-Fil		M = 1375, R = 50	M = 1763, R = 31
Low Means			M = 2094, R = 53

Supervised SCCA [15] was performed on the unfiltered data, after SuMO-Fil, after low mean filtering, and after low variance filtering. This table presents the Hamming distance which counts the number of discrepancies between network selections. The larger the Hamming distance suggests the selections are more different from each other. Let M denote DNA methylation and R denote gene expression.