Table 2.

Maximum Likelihood Parameter Estimates for the Best-Fitting Models for INDELs in CDS Regions and Noncoding regions

Model and DFE	Variant Type	C	θ	γ	Scale	Shape	ϵ	α (%)
CDS: equal mutation rate	Insertions	1	$4.92\times {10}^{-6}$	–1.14	—	—	0.0799
Discrete C = 2	Insertions	2	0.000134	–801	—	—	0.000307	71
Ancestral repeat reference	Deletions	1	$8.32\times {10}^{-6}$	–2.70	—	—	0.0368
	Deletions	2	0.000206	–649	—	—	$3.12\times {10}^{-7}$	86
CDS: equal mutation rate	Insertions	1	$4.79\times {10}^{-6}$	–0.264	—	—	0.0729
Discrete C = 2	Insertions	2	0.000156	–897	—	—	0.000526	63
Noncoding reference	Deletions	1	$7.79\times {10}^{-6}$	–1.70	—	—	0.0366
	Deletions	2	0.000205	–629	—	—	0.00587	79
Noncoding: free mutation rate	Insertions	—	0.000170	–53.6	1,553	0.0345	0.0110	—
Continuous	Deletions	—	0.000293	–75.5	715	0.106	0.0166	—

Note.—C defines the number of site class, θ the population scaled mutation rate, γ the population scaled selection coefficient, ϵ the polarization error, and α the proportion of INDEL substitutions driven by positive selection. Where γ values are presented for the continuous model these are mean γ estimates and the product of the scale and shape parameters.