Position/Title: CGIL Collaborator / Livestock Data Technician
Office: ANNU 110A
Lucas obtained his Bachelor’s in Biotechnology (2015) and a Master’s in Biosciences (2017) from the Federal University of Bahia, Brazil, and recently (June 2022) defended his Ph.D. in Animal Biosciences (Anima Breeding and Genetics) at the University of Guelph.
As part of numerous research projects during his undergraduate studies, he acquired experience in genetic polymorphisms, molecular biology, fermentative processes, bioprospection of microorganisms, and the production and purification of industrial enzymes. He was an exchange student at the University of Toronto (2013), where he improved his skills in molecular biology and enzymology while studying the biological degradation of pine tree bark by different fungi species for bioethanol production.
As a master’s student, he decided to understand the role of different bacteria from the rumen in the degradation of plant cell-wall through bioinformatics analysis and to target species with the potential to produce enzymes of industrial interest. His passion for research and new challenges brought him to the Centre for Genetic Improvement of Livestock (CGIL) at the University of Guelph in 2018 to pursue a Ph.D. in Animal Breeding and Genetics under the advisement of Dr. Flavio Schenkel and Dr. Christine Baes, where he applied machine learning algorithms to improve breeding programs in the dairy industry.
Lucas currently works at the OMAFRA Elora/Ponsonby Livestock Research Centres as a Livestock Data Technician and collaborates with CGIL providing data support for students and researchers.
Previous Research Projects
- Conformation traits of Holstein cows and their association with the Pro$ selection index. There are currently more than 20 conformation traits being genetically evaluated for Canada Holsteins. Understanding the contribution of each of these traits to a monetary index, such as the national selection index Pro$, would provide information to guide the Canadian dairy industry on how to best consider these traits in recording and genetic evaluation programs. This study uses multiple linear regression and principal component analysis to assess the association between conformation traits and Pro$. A video of the oral presentation at the 2020 American Dairy Science Association Meeting with initial results from this research is publicly available here. This work was done under the advisement of Dr. Flavio Schenkel and Dr. Christine Baes, in collaboration with Dr. Gerson Oliveira.
- Classification of breeding code protocol descriptions used with Canadian Holsteins. Dairy farmers are constantly investing in ways to ensure cows become pregnant in an optimal and timely manner. Timed artificial insemination (TAI) has proven to be a successful management tool in dairy cattle, but it is known to mask an animal’s true fertility performance, reducing the accuracy of genetic evaluations for fertility traits. Splitting fertility traits by the management technique involved in the breeding can be a viable technique to address the bias. Nonetheless, there is a lack of specificity and uniformity in the recording of breeding code protocol descriptions (BC) by dairy farmers. Therefore, this project proposed the development of a supervised machine learning model to classify BC into TAI or other protocols, opening the way for unbiased genetic evaluation of animals according to their true fertility performance. This work was done under the advisement of Dr. Flavio Schenkel and Dr. Christine Baes, in collaboration with Colin Lynch, Dr. Gerson Oliveira, and Dr. Dan Tulpan.
- Classification of insemination outcome in historical data of Holstein cattle using decision tree-based ensemble algorithms. Currently, the national Data Exchange System in place in Canada stores insemination records performed by artificial insemination technicians, but those do not include the insemination outcome, which is recorded on the farmers’ herd management software. Knowing the outcome of large volumes of previous inseminations without the need to resort to herd management software from individual farms, but only to a curated and centralized national database is certainly beneficial for researchers and artificial insemination organizations. Therefore, this work developed machine learning supervised classification models able to determine the outcome of previous insemination records of Canadian Holsteins as open, pregnant, or aborted. This work was done under the advisement of Dr. Flavio Schenkel and Dr. Christine Baes, in collaboration with Dr. Gerson Oliveira, and Dr. Dan Tulpan.
Featured Web-Applications Developed
- https://alcantara.shinyapps.io/prodollar - Supplementary material to research paper featuring extra information, such as the complete dataset used for the study, density plots (number of daughters, relative breeding values, and reliabilities), result summaries from regression analysis, AIC values from Stepwise-Backward Regression, full results from PCA (all eigenvalues and eigenvectors), and extra plots from PCA (explained variance, variable contribution, quality of representation of traits, and interactive biplot).
- https://cgil.shinyapps.io/correlations - Supplementary material to research paper featuring an interactive tool to facilitate visualization of genetic parameters of all currently genetically evaluated traits of Canadian Holsteins (correlation matrix, network plot, genetic trends, trait definitions, etc)
- https://alcantara.shinyapps.io/covid - Daily updates of COVID-19 notifications in Brazil with simple interactive graphs [Available in Portuguese only]
- Alcantara, L. M., C.F. Baes, G.A. Oliveira Junior and F.S. Schenkel. 2022. Conformation traits of Holstein cows and their association with a Canadian economic selection index. Canadian Journal of Animal Science. https://doi.org/10.1139/CJAS-2022-0013.
- Oliveira Junior, G.A., F.S. Schenkel, L.M. Alcantara, K. Houlahan, C. Lynch, and C.F. Baes. 2021. Estimated genetic parameters for all genetically evaluated traits in Canadian Holsteins. Journal of Dairy Science. https://doi.org/10.3168/jds.2021-20227.