Comparison of three boosting methods in parent-offspring trios for genotype imputation using simulation study

Table 1 Mean of imputation accuracy for Boosting methods in various versions on the four different datasets

Data set	Density	Sample size	Version	AB	LB	TB
	5 k	100	NA10	0.9843	0.9954	0.9611
	5 k	100	NA30	0.9883	0.9947	0.9638
G1	5 k	100	NA50	0.9822	0.9909	0.9621
	5 k	100	NA70	0.9777	0.9829	0.9583
	5 k	100	NA90	0.9211	0.9303	0.9246
			Mean	0.9707	0.9788	0.9539
	10 k	100	NA10	0.9861	0.9981	0.9702
	10 k	100	NA30	0.9886	0.9978	0.9697
G2	10 k	100	NA50	0.9912	0.9970	0.9679
	10 k	100	NA70	0.9898	0.9939	0.9647
	10 k	100	NA90	0.9653	0.9714	0.9523
			Mean	0.9842	0.9916	0.9649
	5 k	500	NA10	0.9859	0.9967	0.9650
	5 k	500	NA30	0.9885	0.9952	0.9650
G3	5 k	500	NA50	0.9877	0.9926	0.9638
	5 k	500	NA70	0.9800	0.9848	0.9618
	5 k	500	NA90	0.9288	0.9383	0.9362
			Mean	0.9741	0.9815	0.9583
	10 k	500	NA10	0.9787	0.9983	0.9706
	10 k	500	NA30	0.9799	0.9977	0.9692
G4	10 k	500	NA50	0.9830	0.9967	0.9665
	10 k	500	NA70	0.9877	0.9959	0.9634
	10 k	500	NA90	0.9706	0.9767	0.9552
			Mean	0.9799	0.9930	0.9649

NA10: 10 % of genotype is missing per offspring, NA30: 30 % of genotype is missing per offspring, NA50: 50 % of genotype is missing per offspring, NA70: 70 % of genotype is missing per offspring, NA90: 90 % of genotype is missing per offspring, Bold: Mean of different versions in each dataset
AB AdaBoost, LB LogitBoost, TB TotalBoost

ISSN: 2055-0391