Dimensionality of genomic information and its impact on genome-wide associations and variant selection for genomic prediction: a simulation study

Table 3 Estimated sample size based on local polynomial regression using ‘ALL’ and ‘H30’ as benchmark

Scenario	Heritability	%Var	\({\text{SS}}_{{{\text{pol}}}}\)	\({\text{EIGx}}_{{{\text{SSpol}}}}\)
Ne20 Q200	H30 (ALL)	50.1
	H90	50.1	3626	EIG95–98 (1719–3952)
	H99	50.1	1622	EIG90–95 (871–1728)
Ne20 Q2000	H30 (ALL)	3.5
	H90	3.5	3652	EIG95–98 (1674–3854)
	H99	3.5	1599	EIG90–95 (833–1658)
Ne200 Q200	H30 (ALL)	74.5
	H90	74.5	4458	EIG80–90 (2617–5182)
	H99	74.5	3153	EIG80–90 (2598–5146)
Ne200 Q2000	H30 (ALL)	17.0
	H90	17.0	4526	EIG80–90 (2609–5175)
	H99	17.0	3206	EIG80–90 (2612–5172)

Dashed arrows in Fig. 5 correspond to the estimated sample size for each line of the H90 and H99 scenarios, which were obtained using local polynomial regression. These values are referred to as SS_pol
%Var: percentage of variance explained by significantly identified QTN; \({\text{SS}}_{{{\text{pol}}}}\): estimated sample size using local polynomial regression; \({\text{EIGx}}_{{{\text{SSpol}}}}\): EIGx scenario range including \({\text{SS}}_{{{\text{pol}}}}\)

ISSN: 1297-9686