Table 4.

Heritability (h²) and additive genetic variances (${\sigma }_a^2$) explained by different subsets of SNPs from Brahman cow population.

Method	${\sigma }_a^2$	${\sigma }_e^2$	${\sigma }_p^2$	h2	% ${\sigma }_a^2$ of all SNPs
RF400	36.37 (16.96)	563.14 (28.96)	599.51 (27.47)	0.061 (0.028)	48.47
RF1,000	45.99 (20.94)	553.36 (30.64)	599.35 (27.47)	0.077 (0.035)	61.29
RF3,000	67.01 (27.79)	533.00 (33.67)	600.02 (27.63)	0.11 (0.046)	89.30
RF18,453†	71.20 (31.70)	529.11 (36.42)	600.32 (27.64)	0.12 (0.052)	94.88
GBM400	39.99 (17.56)	559.71 (28.98)	599.69 (27.52)	0.067 (0.029)	53.29
GBM1,000	61.87 (23.71)	537.93 (31.02)	599.80 (27.66)	0.10 (0.039)	82.45
GBM3,000	82.07 (30.70)	518.64 (34.60)	600.71 (27.83)	0.14 (0.050)	109.37
GBM16,600†	73.57 (32.25)	526.86 (36.62)	600.37 (27.67)	0.12 (0.053)	98.04
XgBoost400	22.28 (15.75)	578.30 (29.64)	600.57 (27.37)	0.037 (0.026)	30.36
XgBoost1,000	35.01 (20.73)	565.45 (31.41)	600.46 (27.43)	0.058 (0.035)	46.66
XgBoost3,000	40.56 (23.45)	559.66 (32.79)	600.22 (27.42)	0.068 (0.039)	54.05
XgBoost9,122†	65.75 (29.30)	534.18 (34.94)	599.63 (27.59)	0.11 (0.049)	87.62
Even400	20.46 (15.25)	580.42 (29.52)	600.88 (27.39)	0.034 (0.026)	27.27
Even1,000	32.42 (19.82)	568.07 (31.13)	600.48 (27.40)	0.054 (0.033)	43.20
Even3,000	53.85 (25.77)	546.55 (33.18)	600.40 (27.56)	0.090 (0.043)	71.76
All SNPs (38,082)	75.04 (32.43)	525.02 (36.77)	600.06 (27.64)	0.125 (0.054)	100.00

${\sigma }_e^2$ – residual variance; ${\sigma }_p^2$ – total phenotypic variance.

Standard errors are given in parentheses.

RF – Random Forests; GBM – Gradient Boosting Machine, XgBoost – Extreme Gradient Boosting; Even – evenly spaced along the genome; The SNPs with positive variable importance values.