Maltecca, C. | bioRxiv Science

bioRxiv2018

Predicting Growth and Carcass Traits in Swine Using Metagenomic Data and Machine Learning Algorithms

In this paper, we evaluated the power of metagenome measures taken at three time points over the growth test period (weaning, 15 and 22 weeks) to foretell growth and carcass traits in 1039 individuals of a line of crossbred pigs. We measured prediction accuracy as the correlation between actual and predicted phenotypes in a five-fold cross-validation setting. Phenotypic traits measured included live weight measures and carcass composition obtained during the trial as well as at slaughter. We employed a null model excluding microbiome information as a baseline to assess the increase in prediction accuracy stemming from the inclusion of operational taxonomic units (OTU) as predictors. We further contrasted performance of models from the Bayesian alphabet (Bayesian Lasso) as well machine learning approaches (Random Forest and Gradient Boosting) and semi-parametric kernel models (Reproducing Kernel Hilbert space). In most cases, prediction accuracy increased significantly with the inclusion of microbiome data. Accuracy was more substantial with the inclusion of metagenomic information taken at week 15 and 22, with values ranging from approximately 0.30 for loin traits to more than 0.50 for back-fat. Conversely, microbiome composition at weaning resulted in most cases in marginal gains of prediction accuracy, suggesting that later measures might be more useful to include in predictive models. Model choice affected predictions marginally with no clear winner for any model/trait/time point. We, therefore, suggest average prediction across models as a robust strategy in fitting metagenomic information. In conclusion, microbiome composition can effectively be used as a predictor of growth and composition traits, particularly for fatness traits. The inclusion of OTU predictors could potentially be used to promote fast growth of individuals while limiting fat accumulation. Early microbiome measures might not be good predictors of growth and OTU information might be best collected at later life stages. Future research should focus on the inclusion of both microbiome as well as host genome information in predictions, as well as the interaction between the two. Furthermore, the influence of microbiome on feed efficiency as well as carcass and meat quality should be investigated.

genetics↗

bioRxiv2017

A heuristic method to identify runs of homozygosity associated with reduced performance in livestock.

While for the most part genome-wide metrics are currently employed in managing livestock inbreeding, genomic data offer, in principle, the ability to identify functional inbreeding. Here we present a heuristic method to identify haplotypes contained within a run of homozygosity (ROH) associated with reduced performance. Results are presented for simulated and swine data. The algorithm comprises 3 steps. Step 1 scans the genome based on marker windows of decreasing size and identifies ROH genotypes associated with an unfavorable phenotype. Within this stage, multiple aggregation steps reduce the haplotype to the smallest possible length. In step 2, the resulting regions are formally tested for significance with the use of a linear mixed model. Lastly, step 3 removes nested windows. The effect of the unfavorable haplotypes identified and their associated haplotype probabilities for a progeny of a given mating pair or an individual can be used to generate an inbreeding load matrix (ILM). Diagonals of ILM characterize the functional inbreeding load of individual (IIL). We estimated the accuracy of predicting the phenotype based on ILL. We further compared the significance of the regression coefficient for IIL on phenotypes to genome-wide inbreeding metrics. We tested the algorithm using simulated scenarios (n =12) combining different levels of linkage disequilibrium (LD) and number of loci impacting a quantitative trait. Additionally, we investigated 9 traits from two maternal purebred swine lines. In simulated data, as the LD in the population increased the algorithm identified a greater proportion of the true unfavorable ROH effects. For example, the proportion of highly unfavorable true ROH effects identified raised from 32 to 41 % for the low to the high LD scenario. In both simulated and real data the haplotypes identified were contained within a much larger ROH (9.12-12.1 Mb). The IIL prediction accuracy was greater than zero across all scenarios for simulated data (high LD scenario mean (95% confidence interval): 0.49 (0.47-0.52)) and for nearly all swine traits (mean {+/-} SD: 0.17{+/-}0.10). On average across simulated and swine datasets the IIL regression coefficient was more closely related to progeny performance than any genome-wide inbreeding metric. A heuristic method was developed that identified ROH genotypes with reduced performance and characterized the combined effects of ROH genotypes within and across individuals.

genomics↗