Meta-analysis of direct and genomic imprinting heritabilities for growth traits in small ruminants

Abstract

A random-effects meta-analysis was conducted to pool published estimates of maternal and paternal genomic imprinting heritability ($h_{\mathrm{mi}}^2$ and $h_{\mathrm{pi}}^2$) for economically important traits in small ruminants. Ten scientific articles (2017–2024) yielded 96 estimates of maternal and paternal genomic imprinting heritability. Cochran’s Q test and the I² statistic were used to assess the heterogeneity of the studies, while Egger’s test was used to evaluate publication bias. A Baujat plot was generated to quantify each study's contribution to overall heterogeneity (OHT) and the weighted parental genomic imprinting heritability estimates (WHE) for each trait. The weighted coefficients of variation varied from 9.51 % for the Kleiber ratio (KR) to 23.57 % for the average daily gain (ADG). The weighted average $h_{\mathrm{mi}}^2$ ranged between 0.015 for KR and 0.111 for birth weight (BWT), while the mean $h_{\mathrm{pi}}^2$ ranged from 0.015 for ADG to 0.049 for weaning weight (WWT). Additionally, the $h_{\mathrm{mi}}^2$ estimates for BWT and six-month weight (SMWT) as well as the $h_{\mathrm{pi}}^2$ estimates for BWT and KR displayed significant heterogeneity (P < 0.05). The incorporation of genomic imprinting effects into genetic models increased the direct heritability (h²) of BWT by 4.72 %. However, the h² values for the other traits decreased significantly, with a reduction of 33 % for SMWT and up to 93 % for yearling weight (YWT). The I² index revealed high heterogeneity for the $h_{\mathrm{mi}}^2$ estimate of BWT (86.30 %), but moderate heterogeneity for KR (41.5 %) and SMWT (58.10 %). For $h_{\mathrm{pi}}^2$, significant heterogeneity was found in BWT (70.90 %) and KR (61.32 %), whereas SMWT showed moderate heterogeneity (36.47 %). The Baujat plot revealed that the estimate of 0.120 ± 0.023 (Ghafouri-Kesbi et al., 2022b_Data set1) significantly contributed to the OHT of weighted $h_{\mathrm{mi}}^2$ for BWT, WWT, YWT, and ADG, while it had a significant effect on both WHE and OHT concerning the mean value of $h_{\mathrm{pi}}^2$ for SMWT. These findings support the integration of genomic imprinting effects in selection models to improve breeding outcomes.