Sustainability & efficiency

Estimating heritability of deep phenotypes for feed efficiency in dairy cows

---

Abstract

Deep phenotypes of feed efficiency are the inner, meta-mechanisms of feed efficiency which are difficult and expensive to record in practice. Examples are the net energy required for maintenance, maximum feed intake capacity, and (net) efficiency of converting net energy to energy in milk constituents. In earlier studies, deep phenotypes were estimated for a limited number of Holstein-Friesians cows. Consequently, genetic parameters for estimated deep phenotypes could not be estimated as there were insufficient cows for genetic analysis. Therefore, this study aims to evaluate the heritability of estimated deep phenotypes from a larger number of Holstein-Friesian cows. Eight deep phenotypes were estimated for 1944 individual Holstein-Friesian dairy cows with a total of 142,781 weekly measured fat- and protein-corrected milk production (FPCM), dry matter intake (DMI), and bodyweight (BW) records. Deep phenotypes were the input parameters to the mechanistic model LiGAPS-Dairy that simulated the same 142,781 weekly FPCM, DMI, and BW records for 1944 cows. A genetic algorithm, an optimization algorithm based on the principle of natural selection, minimized the difference between simulated results and measured weekly FPCM, DMI, and BW records by changing the value of deep phenotypes, where the deep phenotypes values were the estimated values at the minimized difference. Out of the 1944 cows, 1795 cows with the best fits between simulated results and measured records were included for genetic analysis. Measured deep phenotypes were unavailable, and the heritability was estimated for the estimated eight deep phenotypes. Three multivariate models were fitted: 1) no fixed effects (M_NF), assuming the mechanistic model accounts for these; 2) herd-year-season of first calving as fixed effect (M_F) and 3) herd-year-season of first calving as random effect (M_R). Overall, the heritability estimates for estimated deep phenotypes ranged from 0.04 ± 0.04 to 0.48 ± 0.05 and were highest in M_NF, lowest in M_F, and intermediate in M_R. Compared to M_NF (0.31 ± 0.05), the additive genetic variance reduced more sharply in M_F (0.06 ± 0.03) than in M_R (0.19 ± 0.04). In conclusion, eight deep phenotypes were found to be heritable, and heritability was largely influenced by the genetic model considered. Hence, this study may provide scope to include deep phenotypes in future breeding programs.

Keywords: 2026

How to Cite:

Adhikari, N., Gredler-Grandl, B., Van der Linden, A. & Veerkamp, R., (2026) “Estimating heritability of deep phenotypes for feed efficiency in dairy cows”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285416. doi: https://doi.org/10.31274/wcgalp.23693

Rights: 1