Deep learning (DL) has emerged as a powerful tool to make accurate predictions from complex data such as image, text, or video. However, its ability to predict phenotypic values from molecular data is less well studied. Here, we describe the theoretical foundations of DL and provide a generic code that can be easily modified to suit specific needs. DL comprises a wide variety of algorithms which depend on numerous hyperparameters. Careful optimization of hyperparameter values is critical to avoid overfitting. Among the DL architectures currently tested in genomic prediction, convolutional neural networks (CNNs) seem more promising than multilayer perceptrons (MLPs). A limitation of DL is in interpreting the results. This may not be relevant for genomic prediction in plant or animal breeding but can be critical when deciding the genetic risk to a disease. Although DL technologies are not “plug-and-play”, they are easily implemented using Keras and TensorFlow public software. To illustrate the principles described here, we implemented a Keras-based code in GitHub.

Genomic prediction (GP) is the procedure whereby the genetic merits of untested candidates are predicted using genome wide marker information. Although numerous examples of GP exist in plants and animals, applications to polyploid organisms are still scarce, partly due to limited genome resources and the complexity of this system. Deep learning (DL) techniques comprise a heterogeneous collection of machine learning algorithms that have excelled at many prediction tasks. A potential advantage of DL for GP over standard linear model methods is that DL can potentially take into account all genetic interactions, including dominance and epistasis, which are expected to be of special relevance in most polyploids. In this study, we evaluated the predictive accuracy of linear and DL techniques in two important small fruits or berries: strawberry and blueberry. The two datasets contained a total of 1,358 allopolyploid strawberry (2n=8x=112) and 1,802 autopolyploid blueberry (2n=4x=48) individuals, genotyped for 9,908 and 73,045 single nucleotide polymorphism (SNP) markers, respectively, and phenotyped for five agronomic traits each. DL depends on numerous parameters that influence performance and optimizing hyperparameter values can be a critical step. Here we show that interactions between hyperparameter combinations should be expected and that the number of convolutional filters and regularization in the first layers can have an important effect on model performance. In terms of genomic prediction, we did not find an advantage of DL over linear model methods, except when the epistasis component was important. Linear Bayesian models were better than convolutional neural networks for the full additive architecture, whereas the opposite was observed under strong epistasis. However, by using a parameterization capable of taking into account these non-linear effects, Bayesian linear models can match or exceed the predictive accuracy of DL. A semiautomatic implementation of the DL pipeline is available at https://github.com/lauzingaretti/deepGP/.

Mitigation of greenhouse gas emissions is relevant for reducing the environmental impact of ruminant production. In this study, the rumen microbiome from Holstein cows was characterized through a combination of 16S rRNA gene and shotgun metagenomic sequencing. Methane production (CH₄) and dry matter intake (DMI) were individually measured over 4–6 weeks to calculate the CH₄ yield (CH₄y = CH₄/DMI) per cow. We implemented a combination of clustering, multivariate and mixed model analyses to identify a set of operational taxonomic unit (OTU) jointly associated with CH₄y and the structure of ruminal microbial communities. Three ruminotype clusters (R1, R2 and R3) were identified, and R2 was associated with higher CH₄y. The taxonomic composition on R2 had lower abundance of Succinivibrionaceae and Methanosphaera, and higher abundance of Ruminococcaceae, Christensenellaceae and Lachnospiraceae. Metagenomic data confirmed the lower abundance of Succinivibrionaceae and Methanosphaera in R2 and identified genera (Fibrobacter and unclassified Bacteroidales) not highlighted by metataxonomic analysis. In addition, the functional metagenomic analysis revealed that samples classified in cluster R2 were overrepresented by genes coding for KEGG modules associated with methanogenesis, including a significant relative abundance of the methyl‐coenzyme M reductase enzyme. Based on the cluster assignment, we applied a sparse partial least‐squares discriminant analysis at the taxonomic and functional levels. In addition, we implemented a sPLS regression model using the phenotypic variation of CH₄y. By combining these two approaches, we identified 86 discriminant bacterial OTUs, notably including families linked to CH₄ emission such as Succinivibrionaceae, Ruminococcaceae, Christensenellaceae, Lachnospiraceae and Rikenellaceae. These selected OTUs explained 24% of the CH₄y phenotypic variance, whereas the host genome contribution was ~14%. In summary, we identified rumen microbial biomarkers associated with the methane production of dairy cows; these biomarkers could be used for targeted methane‐reduction selection programmes in the dairy cattle industry provided they are heritable.

Background: The pig gut microbiome harbors thousands of species of archaea, bacteria, viruses and eukaryotes such as protists and fungi. However, since the majority of published studies have been focused on prokaryotes, little is known about the diversity, host-genetic control, and contributions to host performance of the gut eukaryotic counterparts. Here we report the first study that aims at characterizing the diversity and composition of gut commensal eukaryotes in pigs, exploring their putative control by host genetics, and analyzing their association with piglets body weight.Results: Fungi and protists from the faeces of 514 healthy Duroc pigs of two sexes and two different ages were characterized by 18S and ITS ribosomal RNA gene sequencing. The pig gut mycobiota was dominated by yeasts, with a high prevalence and abundance ofKazachstaniaspp. Regarding protists, representatives of four genera (Blastocystis, Neobalantidium, Tetratrichomonas and Trichomitus) were predominant in more than the 80% of the pigs. Heritabilities for the diversity and abundance of gut eukaryotic communities were estimated with the subset of 60d aged piglets (N = 390). The heritabilities of α-diversity and of the abundance of fungal and protists genera were low, ranging from 0.15 to 0.28. A genome wide association study reported genetic variants related to the fungal α-diversity and to the abundance ofBlastocystisspp. Annotated candidate genes were mainly associated with immunity, gut homeostasis and metabolic processes. Additionally, we explored the association of gut commensal eukaryotes with piglet body weight. Our results pointed to a positive contribution of fungi from theKazachstaniagenus, while protists displayed both positive (BlastocystisandEntamoeba) and negative (Trichomitus) associations with piglet body weight.Conclusions: Our results point towards a minor and taxa specific genetic control over the diversity and composition of the pig gut eukaryotic communities. Moreover, we provide evidences of the associations between piglets’ body weight after weaning and members from the gut fungal and protist eukaryote community. Overall, this study highlights the relevance of considering, along with that of bacteria, the contribution of the gut eukaryote communities to better understand host-microbiome association and their role on pig performance, welfare and health.

Background: Analysis and prediction of complex traits using microbiome data combined with host genomic information is a topic of utmost interest. However, numerous questions remain to be answered: how useful can the microbiome be for complex trait prediction? Are estimates of microbiability reliable? Can the underlying biological links between the host’s genome, microbiome, and phenome be recovered? Methods: Here, we address these issues by (i) developing a novel simulation strategy that uses real microbiome and genotype data as inputs, and (ii) using variance-component approaches (Bayesian Reproducing Kernel Hilbert Space (RKHS) and Bayesian variable selection methods (Bayes C)) to quantify the proportion of phenotypic variance explained by the genome and the microbiome. The proposed simulation approach can mimic genetic links between the microbiome and genotype data by a permutation procedure that retains the distributional properties of the data. Results: Using real genotype and rumen microbiota abundances from dairy cattle, simulation results suggest that microbiome data can significantly improve the accuracy of phenotype predictions, regardless of whether some microbiota abundances are under direct genetic control by the host or not. This improvement depends logically on the microbiome being stable over time. Overall, random-effects linear methods appear robust for variance components estimation, in spite of the typically highly leptokurtic distribution of microbiota abundances. The predictive performance of Bayes C was higher but more sensitive to the number of causative effects than RKHS. Accuracy with Bayes C depended, in part, on the number of microorganisms’ taxa that influence the phenotype. Conclusions: While we conclude that, overall, genome-microbiome-links can be characterized using variance component estimates, we are less optimistic about the possibility of identifying the causative host genetic effects that affect microbiota abundances, which would require much larger sample sizes than are typically available for genome-microbiome-phenome studies. The R code to replicate the analyses is in https://github.com/miguelperezenciso/simubiome .