Addressing scope of inference for global genetic evaluation of livestock

Abstract

-Genetic evaluations should become more accurate with the advent of whole genome selection (WGS) based on high density SNP panels. The use of WGS should then accelerate genetic gains for production traits given likely decreases in generation interval due to the greater intent to select more animals based just on their genotypes rather than phenotypes.However, past and current genetic evaluations may not generally connect well to the intended scope of inference. For example, estimating haplotype effects from the data of a single reference population does not bode well for the use of WGS in other diverse environments since the scope of inference is too narrow; conversely, WGS based on estimates, for example, derived from daughter yield deviations of dairy bulls may be too broad to infer upon genetic merit under any one particular environment.The treatment of contemporary group effects as random rather than as fixed, heterogeneous variances, genotype by environment interaction, and multiple trait analyses are all important scope of inference issues that are discussed in this review.Management systems and environments have and will continue to change; hence, it is vital that genetic evaluations are as robust and scope-appropriate as is possible in order to optimize animal adaptation to these changes.Key Words: Bayesian inference, genotype by environment interaction, heterogeneous variances, mixed effects models, multiple traits, random effects Direcionamento do âmbito de inferência para avaliação genética global dos animais RESUMO -As avaliações genéticas devem se tornar mais precisas com o advento da seleção pelo genoma inteiro (WGS), com base em painéis SNP de alta densidade. O uso da WGS deve, portanto, acelerar os ganhos genéticos para características de produção, em razão da provável diminuição do intervalo de geração, pela maior intenção de selecionar animais com base apenas nos seus genótipos, em vez de fenótipos. Contudo, as avaliações genéticas passadas e atuais podem, de forma geral, não se adaptar bem com o âmbito de inferência pretendida. Por exemplo, estimar efeitos de haplótipo a partir do uso de uma única população de referência não traz boa predição para uso geral da WGS em outros ambientes diversos, já que o âmbito de inferência é muito restrito; contrariamente, WGS baseada em estimativas derivadas dos desvios de produção das filhas de touros leiteiros podem ser muito generalizadas para inferir sobre o mérito genético em qualquer ambiente em particular. O tratamento dos efeitos de grupo contemporâneo, como fixo ou aleatório, variâncias heterogêneas, interação genótipo-ambiente e as análises de características múltiplas são questões importantes no âmbito da inferência e são discutidas nesta revisão. Os sistemas de manejo e os ambientes têm mudado e continuarão a mudar; deste modo, é essencial que avaliações genéticas sejam tão robustas e adequadas ao contexto quanto possível, a fim de otimizar a adaptação dos animais a essas mudanças.Palavras-chave: características múltiplas, efeitos aleatór...

Invited review: Reproducible research from noisy data: Revisiting key statistical principles for the animal sciences

“…In this context, scope of inference also can be framed as an issue of herd × treatment interactions driven by herd-specific treatment effects. This is analogous to the concept of genotype-by-environment interaction in quantitative genetics (Tempelman, 2010) or study-by-treatment interaction within the context of the increasingly popular meta-analysis (St-Pierre, 2001).…”

“…This is a worthwhile consideration, particularly under the framework of systems biology (Sinoquet, 2014), whereby multiple levels of organization may be simultaneously at play. Levels might range from fine-grained (e.g., molecules or cells) to physiological mechanisms in whole living organisms (Sinoquet, 2014), or from individual animals to management clusters (e.g., pens or herds) to a whole livestock production industry across national or even international landscapes (Tempelman, 2010). In other words, scope of inference frames the inherently multilevel hierarchy of biological systems and thus outlines the expected reach of results.…”

“…In principle, it is the highest level of organization represented in a data set that ultimately outlines the reference population represented and, thus, the largest possible scope of inference for any conclusion drawn from the study (Mclean et al, 1991;Robinson, 1991). For instance, the inferential space for a study may be narrowly defined and extremely local, limited to a single animal (e.g., if data were obtained from multiple tissues sampled from that individual), to a manage-ment-defined group of animals (e.g., pen), or even to a single herd (e.g., if multiple animals from that single herd were sampled; Tempelman, 2009Tempelman, , 2010). The latter is a common occurrence in the animal sciences, as many studies are conducted within the confines of an individual herd, be it a research site or a commercial operation.…”

“…Even if considered typical of the population, the single selected herd grants no true replication from such population and, as a result, the inherent variability among herds cannot be estimated. Data collected from a single herd delineates a strictly narrow inferential space (Mclean et al, 1991;Tempelman, 2010) limited to just that herd. Admittedly, single-herd studies may serve as pilot and provide preliminary evidence for further multiherd experimental studies or even to guide exploration of observational field data from commercial livestock operations (Rosa and Valente, 2013).…”

“…In general terms, specifying blocks as random effects allows for the recovery of interblock information, thus enhancing inferential efficiency (Yates, 1940). Random blocks also allow for a broader scope of inference of treatments of interest beyond the blocks specifically involved in a study, encompassing the population of blocks from which those enrolled in a study are presumably representative (Tempelman, 2010). Instead, a fixed effects specification of blocks (i.e., herds, pens) formally impedes such broad interpretation, despite any generalizability intended by the researcher (Mclean et al, 1991).…”