EnvRtype: a software to interplay enviromics and quantitative genomics in agriculture

Abstract

Envirotyping is an essential technique used to unfold the non-genetic drivers associated with the phenotypic adaptation of living organisms. Here we introduce the EnvRtype R package, a novel toolkit developed to interplay large-scale envirotyping data (enviromics) into quantitative genomics. To start a user-friendly envirotyping pipeline, this package offers: (1) remote sensing tools for collecting (get_weather and extract_GIS functions) and processing ecophysiological variables (processWTH function) from raw environmental data at single locations or worldwide; (2) environmental characterization by typing environments and profiling descriptors of environmental quality (env_typing function), in addition to gathering environmental covariables as quantitative descriptors for predictive purposes (W_matrix function); and (3) identification of environmental similarity that can be used as an enviromic-based kernel (env_typing function) in whole-genome prediction (GP), aimed at increasing ecophysiological knowledge in genomic best-unbiased predictions (GBLUP) and emulating reaction norm effects (get_kernel and kernel_model functions). We highlight literature mining concepts in fine-tuning envirotyping parameters for each plant species and target growing environments. We show that envirotyping for predictive breeding collects raw data and processes it in an eco-physiologically-smart way. Examples of its use for creating global-scale envirotyping networks and integrating reaction-norm modeling in GP are also outlined. We conclude that EnvRtype provides a cost-effective envirotyping pipeline capable of providing high quality enviromic data for a diverse set of genomic-based studies, especially for increasing accuracy in GP across untested growing environments.

Enviromic assembly increases accuracy and reduces costs of the genomic prediction for yield plasticity

“…However, most of those objectives will be hampered if the MET-GP platforms do not consider models with a higher biological meaning (Hammer et al, 2019) and reliable environmental information. A cost-effective solution for that, if the breeder has no access to sensor network tools, relies on the use of remote sensing tools to collect and process basic weather and soil data, such as those available in the EnvRtype R package (Costa-Neto et al, 2021b).…”

“…In general, studies involving FR analysis found that the effect of high temperatures at grain-filling and maturation (Epinat- Le Signor et al, 2001;Romay et al, 2010) , water balance at flowering (Epinat-Le Signor et al, 2001;Millet et al, 2019) and intercept radiation at the vegetative phase (Millet et al, 2019) are the main drivers of G×E for yield components in maize. Thus, Millet et al (2019) From the aspects mentioned above, we envisage that the use of GP for multi-environment predictions must account for some degree of ecophysiological reality while also considering the balance and the relation between parsimony and accuracy (Hammer et al, 2019;Costa-Neto et al, 2021b;Cooper et al, 2021). Here we also highlight in our literature review that multi-environment GP must account for the impact of (1) resource availability in the creation of biologically accurate platforms in training CGM-based approaches and delivering reliable envirotyping information for those purposes, (2) availability of the knowledge of experts in training CGM approaches.…”

“…The presence of both genotype-envirotype and envirotype-phenotype covariances might explain the gains in the predictive ability due to the use of multi-environment GP models in contrast to singleenvironment GP models (Bandeira e Souza et de Oliveira et al, 2020) and why deep learning approaches have successfully captured intrinsic G×E patterns and translated them into gains in accuracy (Montesinos-López et al, 2018;Crossa et al, 2019;Cuevas et al, 2019) . Conversely, this also might explain the need to incorporate secondary sources of information in the prediction of grain yields across multiple environments (Westhues et al, 2017;Ly et al, 2018;Millet et al, 2019;Costa-Neto et al, 2021a;2021b;Jarquín et al, 2020) , as well as the possible limitations of CGM approaches contrasting scenarios differing from those targeted near-iso conditions of CGM calibration (e.g., Cooper et al, 2016;Messina et al, 2018). Thus, an alternative can be supervised approaches to describe the environmental relatedness, such as in this paper, and perhaps unsupervised algorithms capable of taking advantage of the covariances related to the genotype-phenotype, genotypeenvirotype, and envirotype-phenotype dynamics.…”

“…Environmental changing scenarios challenge agricultural research to deliver climate-smart solutions in a time-reduced and cost-effective manner (Tigchelaar et al, 2018;Ramírez-Villegas et al 2020;Cortés et al, 2020). Characterizing crop growth conditions is crucial for this purpose (Xu, 2016) , allowing a deeper understanding of how the environment shapes past, present, and future phenotypic variations (e.g., Ramírez-Villegas et al 2018;Heinemann et al, 2019;Cooper et al, 2014;de los Campos et al, 2020;Costa-Neto et al, 2021b;Antolin et al, 2021). For plant breeding research, mostly based on selecting the best-evaluated genotypes for a target population of environments (TPE), this approach is useful to discriminate genomic and non-genomic sources of crop adaptation.…”

“…Therefore, novel ways to include environmental data (Heslot et al, 2014;Jarquín et al, 2014;Ly et al, 2018;Millet et al, 2019;Gillberg et al, 2019;Costa-Neto et al, 2021a) and process-based crop growth models (CGM) (Messina et al, 2018;Toda et al 2020;Robert et al, 2020;Cooper et al, 2021) in GP are considered the best pathways to fix it. Most of the success achieved by such approaches lies in a better understanding of the visible ecophysiology interplay between genomics and environment variation (Gage et al, 2017;Li et al, 2018;Guo et al, 2020;Costa-Neto et al, 2021b).…”