Aarhus University Seal / Aarhus Universitets segl

GENETIC PARAMETERS FOR FEED EFFICIENCY USING 3D CAMERAS IN JERSEY COWS IN COMMERCIAL DANISH FARMS

Publikation: Bidrag til bog/antologi/rapport/proceedingKonferencebidrag i proceedingsForskningpeer review

The biggest challenge of including feed intake in the breeding goal is to have enough amount of records, given its difficulty measuring individually. However, with new tools available, like 3D cameras and artificial intelligence, this problem might be overcome. This is a preliminary work on getting genetic parameters for dry matter intake (DMI) and body weight (BW) measured using 3D cameras and artificial intelligence to posteriorly calculate residual feed intake (RFI). A total of 21,386 weekly records of DMI and BW recorded from 3D cameras during 2019 and 2020 were available from 505 commercial Danish Jersey cows. These weekly records were complemented with milk and milk content records of the same period, and energy corrected milk (ECM) was calculated. RFI was calculated as the partial regression of dry matter intake on energy sinks (Tempelman et al., 2015). Estimated heritabilities were 0.07 (RFI), 0.08 (DMI), 0.22 (BW) and 0.26 (ECM). Genetic correlations between DMI and ECM (0.23) and BW (0.22) were positive and moderate. Genetic correlations of RFI and DMI were highly positive (0.89), whereas between RFI and BW (0.15) and ECM (-0.20) were low, but given the large standard error not different from zero. Phenotypic correlations between RFI and ECM, BW were close to zero as expected, whereas, between RFI and DMI were close to one. With these results, we conclude that feed efficiency (RFI) calculated using DMI and BW measured by 3D cameras is heritable. However, there still room for improvement. Given that this is a preliminary result, and DMI and BW were measured only in two commercial farms with a limited number of animals, therefore, when adding more farms, animals and records the genetic parameters for DMI, BW and RFI might change. In addition, the algorithm to predict the phenotype with the 3D images still in adjusting phase.
OriginalsprogEngelsk
TitelAAABG Proceedings 2021
DOI
StatusAccepteret/In press - 16 feb. 2021

Se relationer på Aarhus Universitet Citationsformater

ID: 218372441