Table 3.

Model comparisons between the proposed and existing methods to classify inbreeding into new and old classes using pedigree and genomic (ROH segments) information

Traita	Inbreedingb		Criterionc
			Adj.R2	RMSE	AIC	BIC
Using Pedigree information
BW	Existing	Kalinowski	0.1561	4.1822	4294.78	4433.10
		P_5L	0.1553	4.1841	4295.45	4433.78
	Proposed	P_10P	0.1580	4.1776	4293.12	4431.44
WW	Existing	Kalinowski	0.5418	22.6225	6741.76	6884.40
		P_5L	0.5402	22.6625	6744.36	6887.00
	Proposed	P_10P	0.5436	22.5781	6738.87	6881.51
YW	Existing	Kalinowski	0.8448	31.3207	6744.06	6884.56
		P_5L	0.8441	31.3937	6747.26	6887.76
	Proposed	P_10P	0.8449	31.3088	6743.54	6884.04
ADG	Existing	Kalinowski	0.8427	0.1366	− 726.03	− 590.06
		P_5L	0.8437	0.1361	− 730.21	− 594.24
	Proposed	P_11P	0.8441	0.1359	− 732.16	− 596.19
Using Genomic information
BW	Existing	ROH_Mclust	0.15583	4.1829	4295.01	4433.34
		ROH_5L	0.15566	4.1833	4295.16	4433.48
	Proposed	ROH_9P	0.15588	4.1828	4294.97	4433.29
WW	Existing	ROH_Mclust	0.53964	22.6759	6745.23	6887.87
		ROH_5L	0.53946	22.6802	6745.52	6888.15
	Proposed	ROH_13P	0.53962	22.6762	6745.25	6887.89
YW	Existing	ROH_Mclust	0.84299	31.5019	6751.99	6892.49
		ROH_5L	0.84304	31.4976	6751.80	6892.30
	Proposed	ROH_7P	0.84299	31.5025	6752.01	6892.52
ADG	Existing	ROH_Mclust	0.84202	0.1369	−723.05	− 587.08
		ROH_5L	0.84205	0.1368	−723.15	−587.18
	Proposed	ROH_7P	0.84211	0.1368	−723.43	−587.46

^aBW birth weight (kg), WW weaning weight (kg), YW yearling weight (kg), ADG average daily gain (kg/d)

^bKalinowski classification of inbreeding based on Kalinowski et al. (2000), P_5L classification of inbreeding based on 5 generations according to literature, P_10P and P_11P classification of inbreeding based on 10 and 11 generations, respectively, as determined by the proposed method, ROH_Mclust classification of inbreeding based on model-based clustering method, ROH_5L classification of inbreeding based on 5 Mb according to literature, ROH_7P ROH_9P, and ROH_13P classification of inbreeding based on 7, 9, and 13 Mb, respectively, as determined by the proposed method

^cAdj.R2 Adjusted R-squared, RMSE Root Mean Squared Error, AIC Akaike’s Information Criteria, BIC Bayesian information criteria