Feed accounts for the largest part of operating cost in dairy cattle production. Improving feed efficiency (FE) of dairy cattle contributes to dairy profits and reduced environmental footprint of dairy production. Feed efficiency phenotypes are generally difficult and costly to measure in a large scale, limiting the accuracy of genetic improvement. This PhD project aims to study the possibility of improving dairy feed efficiency by integrating genomic selection and phenotype prediction for feed efficiency using data-driven technologies including machine learning.
The successful student will analyse individual cow records from the SRUC’s world-leading Langhill research herd, currently in Dumfries, Scotland. The herd has been continuously operational for nearly five decades since the 1970s, accumulating a world-leading and genetically unique database including longitudinal data for milk production, fertility, health, welfare, and feed efficiency. The research herd has won SRUC Queen's Anniversary Prize and has been widely recognized as a unique resource for dairy genetics research whose feed efficiency data will be available for the current PhD project.
Three studies are included in the PhD project:
(i) Phenotype prediction for dairy FE using routinely available milk infrared spectra data together with data on milk production, milk composition, and body weight. Prediction methods of partial least square (PLS) regression and deep learning neural network (DLNN) will be used to evaluate and compare the prediction abilities.
(ii) Genome-wide association studies for FE phenotypes at different lactation stages will be carried out to detect large-effect markers associated with FE and changes in genetic background of FE across lactation.
(iii) Improving genomic prediction for FE using predicted records from (i) and genomic results from (ii). Methods of combining actual FE records and predicted records by introducing weights into prediction will be examined. The best ratio of actual records to predicted records in genomic prediction for FE will be determined. In addition, large-effect markers associated with FE from (ii) will be used as biological priors to improve prediction accuracy for FE.
The project integrates data-driven technologies (machine learning, genomic selection) into dairy genetic research. Relevant specialised training will be available in the PhD project.
Application Process:
Please visit this page for full application instructions http://www.eastscotbiodtp.ac.uk/how-apply-0
1) Download and complete the Equality, Diversity and Inclusion survey.
2) Download and complete the EASTBIO Application Form.
3) Submit both to SRUC, [Email Address Removed].
Completed applications must include the following documents:
- Completed EASTBIO application form
- 2 References (to be completed on the EASTBIO Reference Form, also found on the EASTBIO website)
- Academic Qualifications
- Equality, Diversity and Inclusion survey
Unfortunately due to workload constraints, we cannot consider incomplete applications. Please make sure your application is complete by the 16th December 2021. Please ask your referees to submit your references directly to [Email Address Removed].
We anticipate that our first set of interviews will be held 7th – 11th February 2022 with awards made from 18 February.