Table 3.

Comparison of four QTL mapping methods and their packages

Case	GCIM-random	GCIM-fixed	ICIM	CIM
Model	Multi-locus model	Multi-locus model	Single-locus model	Single-locus model
Model transformation	FASTmrEMMA algorithm	FASTmrEMMA algorithm	NA	Interval mapping for ${y^{\prime }}_i=y_i-\sum _{k\ne l,l+1}(x_{ik}{a}_k+z_{ik}{d}_k)$
QTL effect	Random	Fixed	Fixed	Fixed
Estimation of QTL effect	REML or ML	REML or ML	ML	ML
Polygenic background control	Polygenic additive and dominant variances via mixed model framework of GWAS	Polygenic additive and dominant variances via mixed model framework of GWAS	The associated markers (cofactors), except the two markers flanking the current mapping interval; their effects are estimated at each position of genome scanning	The cofactors except for the two markers flanking the current mapping interval; the effects for all the cofactors are estimated only one time
No. of variance components	Five	Three	NA	NA
Polygenic-to-residual variance ratio	Fixed	Fixed	NA	NA
Running time	Fast	Fast	Fast	Slow
Software	GCIM-random and GCIM-fixed: QTL.gCIMapping (https://cran.r-project.org/web/packages/QTL.gCIMapping/index.html) QTL.gCIMapping.GUI (https://cran.r-project.org/web/packages/QTL.gCIMapping.GUI/index.html)
	ICIM: QTL IciMapping (http://www.isbreeding.net/)
	CIM: Windows QTL Cartographer (https://brcwebportal.cos.ncsu.edu/qtlcart/WQTLCart.htm)