Table 2.

Logistic Regression for predicting transmission based on number of glycosites in the variable loops

	Univariate Odds Ratio (95% confidence interval)	p-value		Odds Ratio with subtype interaction (95% confidence interval)	p-value	Test for Homogeneity between subtype, p-value	Wilcoxon Rank Sum p-value
PNGs in V1 loop			PNGs in V1 loop
1 to 2			Subtype B 1 to 2				0.109
3	1.11 (0.59, 2.06)	0.751	3	1.42 (0.55, 3.62)	0.466
≥ 4	1.62 (0.83, 3.14)	0.155	≥ 4	2.29 (0.83, 6.36)	0.111
			Subtype C 1 to 2				0.931
			3	0.66 (0.18, 2.48)	0.541
			≥ 4	0.54 (0.14, 2.10)	0.371	0.347
PNGs in V2 loop			PNGs in V2 loop
0 to 1			Subtype B 0 to 1				0.533
2 to 3	0.75 (0.44, 1.27)	0.282	2 to 3	0.84 (0.44, 1.63)	0.615
			Subtype C 0 to 1				0.323
			2 to 3	0.67 (0.22, 2.13)	0.506	0.509
PNGs in V3 loop			PNGs in V3 loop	*
0			Subtype B 0				0.024
1	2.56 (0.49, 13.43)	0.267	1
			Subtype C 0				0.155
			1			0.008
PNGs in V4 loop			PNGs in V4 loop
0 to 1			Subtype B 0 to 1				0.572
2	0.51 (0.23, 1.14)	0.102	2	0.36 (0.10, 1.31)	0.122
≥ 3	0.54 (0.25, 1.20)	0.133	≥ 3	0.46 (0.13, 1.60)	0.219
			Subtype C 0 to 1				0.329
			2	1.83 (0.35, 9.60)	0.473
			≥ 3	1.19 (0.22, 6.28)	0.842	0.001

Note: In groups where no sequences have 0 glycosites, logistic regression cannot be performed.