Naslov
Ancestral inbreeding and inbreeding-purging models’ comparative analysis based on their classification efficiency of the Pannon white rabbit kits’ survival at birth

Autori
Nguyen, T. ; Nagy, Istvan ; Farkas, J. ; Curik, Ino ; Kövér, G.

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
Book of Abstracts of 1st EAAP Regional Meeting of The European Federation of Animal Science / - , 2023, 90-90

Skup
1st EAAP Regional Meeting of The European Federation of Animal Science

Mjesto i datum
Nitra, Slovačka, 26.04.2023. - 28.04.2023

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Inbreeding-purging models ; Pannon white rabbits ; Survival at birth

Sažetak
The authors analysed 22718 kindling records of 5166 rabbits born between 1992 and 2017, respectively. The 25-year-long exanimated period was separated into three different sub-periods (1992-1997 ; 1997-2007 and 207-2017). The effect of dam and litter inbreeding on the survival of rabbit kits at birth was evaluated with two fundamentally different methods with the same purpose of signalling the possibility of purging. The first method considered the different ancestral inbreeding coeffi cients (AINB) by applying the generalized linear mixed model extended with model averaging. The inbreeding- purging model (IPM) was based on the purged inbreeding coefficient where the possible purging is accounted for. The parity and season of kindling effects were the same in both methods. The classification abilities of these models were compared based on the ROC Curve where the greater area under the curve (AUC) signals better classification performance. The AUC values of the models using AINB ranged between 0.79 and 0.83. On the contrary, IPM resulted in lower AUC values ranging between 0.54 and 0.57. The optimal threshold was determined by the maximum value of Kappa and depending on the examined period they ranged between 0.87-0.89 for the AINB models and it was 0.94 for the IPM. True positive rate (sensitivity) was high both for AINB-based models (0.93-0.96) and for IPM (0.71-0.82). On the contrary true negative rates (specificity) were generally much lower both for the AINB models (0.25-0.38) and for the IPM (0.23-0.37). Consequently, the proportion of the correctly classified cases were substantially higher (0.91- 0.92) for the AINB models compared to the IPM (0.71-0.82).

Izvorni jezik
Engleski

Znanstvena područja
Poljoprivreda (agronomija)

POVEZANOST RADA