Metagenomic Predictions of Growth and Carcass Traits in Pigs with the Use of Bayesian Alphabet and Machine Learning Methods.

C. Maltecca, D. Lu, F. Tiezzi, C. Schillebeeckx, N.P. McNulty, C. Schwab and C. Shull North Carolina State University, Raleigh, USA, Matatu Inc., Saint Louis, MO, USA, The Maschhoffs LLC., Carlyle, IL, USA. Summary In this paper, we evaluated the power of metagenome measures taken at three time points over the growth test period (weaning ,15 weeks and 22 weeks) to predict growth and carcass traits in a line of crossbred pigs. Models from the Bayesian alphabet (Bayesian Lasso) as well as two machine learning approaches (Random Forest and Gradient Boosting) were employed to predict weight, backfat, loin depth and loin area at week 15 and 22. Prediction accuracy was measured as correlation between true and predicted phenotypes in cross validation. In addition, a time dependent recurrent neural network using all microbiome measures simultaneously was fitted to classify individuals in 4 groups based on daily gain and backfat at week 22. In most cases prediction accuracy increased with the inclusion of microbiome composition. Accuracy was larger with the inclusion of microbiome composition taken at week 15 and 22, with values ranging from ~.30 for loin traits to > .50 for backfat. Model choice only affected prediction accuracy marginally. Microbiome can be used as an effective tool to predict growth and carcass in swine.

[1]  C. Maltecca,et al.  The relationship between different measures of feed efficiency and feeding behavior traits in Duroc pigs. , 2017, Journal of animal science.

[2]  Steven Salzberg,et al.  BIOINFORMATICS ORIGINAL PAPER , 2004 .

[3]  Yoshua Bengio,et al.  Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling , 2014, ArXiv.

[4]  T. Tribout,et al.  Optimized management of genetic variability in selected pig populations. , 2008, Journal of animal breeding and genetics = Zeitschrift fur Tierzuchtung und Zuchtungsbiologie.

[5]  Genome-wide association study on legendre random regression coefficients for the growth and feed intake trajectory on Duroc Boars , 2015, BMC Genetics.

[6]  Keiichi Suzuki,et al.  Genetic parameters for measures of residual feed intake and growth traits in seven generations of Duroc pigs , 2009 .

[7]  C. Maltecca,et al.  Feed intake, average daily gain, feed efficiency, and real-time ultrasound traits in Duroc pigs: II. Genomewide association. , 2014, Journal of animal science.

[8]  M. Pop,et al.  Metagenomic Analysis of the Human Distal Gut Microbiome , 2006, Science.

[9]  James Versalovic,et al.  Human microbiome in health and disease. , 2012, Annual review of pathology.

[10]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[11]  Jo Handelsman,et al.  Toward a Census of Bacteria in Soil , 2006, PLoS Comput. Biol..

[12]  G. de los Campos,et al.  Genome-Wide Regression and Prediction with the BGLR Statistical Package , 2014, Genetics.

[13]  John J. McGlone,et al.  Pig Production: Biological Principles and Applications , 2002 .

[14]  P. Somervuo,et al.  Quality Control and Preprocessing , 2014 .