. 2016 May 12;11(5):e0155270. doi: 10.1371/journal.pone.0155270

Table 3. Conjoint analysis of the effects of individual tSNPs on EC risk by CART.

Terminal nodes	Genotype of participants in each node	Cases (%)	Controls (%)	aOR (95% CI)^a	P^a
2	TGFBR1 rs6478974 (TA+AA)	234 (35.7)	421 (64.3)	Reference
6	TGFBR1 rs6478974 (TT) +TGFBR1 rs10512263 (TT) +	134 (40.7)	195 (59.3)	1.13 (0.81–1.59)	0.4683
6	TGFB1 rs4803455 (CA+AA)	134 (40.7)	195 (59.3)	1.13 (0.81–1.59)	0.4683
8	TGFBR1 rs6478974 (TT) +TGFBR1 rs10512263 (TT) +	35 (40.7)	51 (59.3)	1.07 (0.61–1.88)	0.8172
8	TGFB1 rs4803455 (CC) +TGFBR1 rs10733710 (GA+AA)	35 (40.7)	51 (59.3)	1.07 (0.61–1.88)	0.8172
7	TGFBR1 rs6478974 (TT) +TGFBR1 rs10512263 (TT) +	63 (66.3)	32 (33.7)	3.71 (2.14–6.43)	<0.0001
7	TGFB1 rs4803455 (CC) +TGFBR1 rs10733710 (GG)	63 (66.3)	32 (33.7)	3.71 (2.14–6.43)	<0.0001
4	TGFBR1 rs6478974 (TT) +TGFBR1 rs10512263 (TC+CC)	50 (86.2)	8 (13.8)	7.86 (3.42–18.07)	<0.0001

tSNPs, tagging single nucleotide polymorphisms; EC, endometrial cancer; CART, classification and regression tree; OR, odds ratios; CI, confidence intervals.

^a Adjusted for BMI, age at menarche, age at primiparity, number of childbirth, menopause status and family history of cancer in first-degree relatives.

Bold numbers denote a statistical significance at 0.05 level.