REVIEW
published: 09 February 2022
doi: 10.3389/fgene.2022.832153
Genomic Selection: A Tool for
Accelerating the Efficiency of
Molecular Breeding for Development
of Climate-Resilient Crops
Neeraj Budhlakoti 1†, Amar Kant Kushwaha2†, Anil Rai1, K K Chaturvedi1, Anuj Kumar1,
Anjan Kumar Pradhan3, Uttam Kumar4, Rajeev Ranjan Kumar1, Philomin Juliana4,
D C Mishra1* and Sundeep Kumar3*
1ICAR- Indian Agricultural Statistics Research Institute, New Delhi, India, 2ICAR- Central Institute for Subtropical Horticulture,
Lucknow, India, 3ICAR- National Bureau of Plant Genetic Resources, New Delhi, India, 4Borlaug Institute for South Asia (BISA),
Ludhiana, India
Edited by:
Vijay Gahlaut, Since the inception of the theory and conceptual framework of genomic selection (GS),
Institute of Himalayan Bioresource extensive research has been done on evaluating its efficiency for utilization in crop
Technology (CSIR), India
improvement. Though, the marker-assisted selection has proven its potential for
Reviewed by:
Aditya Pratap, improvement of qualitative traits controlled by one to few genes with large effects. Its
Indian Institute of Pulses Research role in improving quantitative traits controlled by several genes with small effects is limited.
(ICAR), India
Upendra Kumar, In this regard, GS that utilizes genomic-estimated breeding values of individuals obtained
Chaudhary Charan Singh Haryana from genome-wide markers to choose candidates for the next breeding cycle is a powerful
Agricultural University, India approach to improve quantitative traits. In the last two decades, GS has been widely
*Correspondence: adopted in animal breeding programs globally because of its potential to improve selection
D C Mishra
dwij.mishra@gmail.com accuracy, minimize phenotyping, reduce cycle time, and increase genetic gains. In
Sundeep Kumar addition, given the promising initial evaluation outcomes of GS for the improvement of
sundeep.kumar@icar.gov.in
yield, biotic and abiotic stress tolerance, and quality in cereal crops like wheat, maize, and
†These authors have contributed
equally to this work rice, prospects of integrating it in breeding crops are also being explored. Improved
statistical models that leverage the genomic information to increase the prediction
Specialty section: accuracies are critical for the effectiveness of GS-enabled breeding programs. Study
This article was submitted to on genetic architecture under drought and heat stress helps in developing production
Plant Genomics,
a section of the journal markers that can significantly accelerate the development of stress-resilient crop varieties
Frontiers in Genetics through GS. This review focuses on the transition from traditional selection methods to GS,
Received: 09 December 2021 underlying statistical methods and tools used for this purpose, current status of GS studies
Accepted: 10 January 2022
Published: 09 February 2022 in crop plants, and perspectives for its successful implementation in the development of
Citation: climate-resilient crops.
Budhlakoti N, Kushwaha AK, Rai A,
Chaturvedi KK, Kumar A, Pradhan AK, Keywords: GS, climate change, STGS, MTGS, abiotic stress, biotic stress, GEBV, climate-resilient crops
Kumar U, Kumar RR, Juliana P,
Mishra DC and Kumar S (2022) INTRODUCTION
Genomic Selection: A Tool for
Accelerating the Efficiency of Molecular
Sustainable food production is the utmost requirement for food and nutritional security. Based on
Breeding for Development of Climate-
Resilient Crops. reports, 821 million people are point below nourishment level; i.e., 151 million children under 5 years
Front. Genet. 13:832153. are stunted; in terms of micronutrients, two billion people are not able to meet the requirement for
doi: 10.3389/fgene.2022.832153 living a healthy life, globally. To meet these demands, the production and supply system has to be
Frontiers in Genetics | www.frontiersin.org 1 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
sound. It has been projected that production has to be increased
by 60% by 2050, amid different challenges related to the
production system posed by climate change (WHO/FAO,
2015), which is further projected to worsen by an increase in
the price of food to the extent of 1–29% by 2050. The
development of climate-resilient varieties through conventional
approaches of hybridization and selection is input-intensive
(labor, land, and time), limiting the realized genetic gain.
Improvement in the genetic gain as per the Lush equation
(Lush, 1943) can be secured through i) better intensity of
selection via accurate and high-throughput phenotyping and
ii) having a broad genetic base representing diverse eco-
geography in breeding program. The advancement in
genomics approaches leads to the availability of huge resources
like genome sequence information, transcriptome, and proteome
that have paved the way to hasten the identification of target
genes mitigating the effects of climate change (Varshney et al., FIGURE 1 | Basic schema of the genomic selection process.
2018). This sequence of information also leads to the
identification of several mutant loci at the nucleotide level
which might be associated with characters of complex nature new selection tool called genomic selection (GS) was proposed
like yield in general and under different circumstances of stress, that can facilitate selection for such traits, by means of net genetic
which are otherwise very difficult to decipher. Genomic selection merit of an individual obtained using the effects of dense markers
emerged as an important tool which can utilize such information distributed across the genome (Meuwissen et al., 2001). In this
for modeling the crop yield for effective and rapid selection under approach, the individual effect of each marker is estimated, and
different environmental conditions to meet the production the additive sum of all the marker effects is used for calculation of
challenges in a climate-changing world. the genomic-estimated breeding values (GEBV) of each
Changes brought about by climate change have affected the individual. In the current scenario of climate change, GS is a
phenology of different crop species leading to a detrimental effect promising tool for improving the genetic gain of individuals
on production and productivity. Different stresses, viz., heat, cold, under the breeding program (Yuan et al., 2019). The basic process
drought, and flood, are specific manifestations of climate change. of any genomic selection process starts with the creation of
Genetic improvement of crops based on phenotypic selection has training population, i.e., individuals having both genotypic and
been successfully achieved through traditional breeding. phenotypic information, and this information is used to build a
However, in recent past, genomics led to the identification of model, where the phenotype is used as a response and genotype as
several underlying genes/QTLs providing tolerance to these a predictor. The information from the developed model is later
specific conditions, which have been utilized in marker- used to estimate the GEBV of breeding population,
assisted selection (MAS). MAS is an indirect selection process, i.e., individuals having only genotypic information. The basic
where individuals for a particular trait of interest are selected process of GS is also explained in Figure 1.
based on the known markers linked to it (Fernando and The major advantage of using GS is that it allows for a drastic
Grossman, 1989). This method has been efficiently used in the reduction in the duration of the breeding cycle as compared to
past for selection of individuals in plant breeding to increase the traditional breeding and also minimizes the cost associated with
selection accuracy compared to the traditional phenotype-based extensive phenotyping, thereby subsequently accelerating genetic
selection process (Mohan et al., 1997). In cereals, MAS resulted in gains and ensuring food and nutritional security (Heffner et al.,
a number of varieties, viz., Improved Pusa Basmati1 2010). However, there are certain factors such as the size of
(Gopalakrishnan et al., 2008), Pusa Basmati 1728 (Singh et al., training and breeding populations, genetic diversity of breeding
2017a), Pusa Basmati 1637 (Singh et al., 2017b), Pusa Samba 1850 population, heritability of the underlying trait, influence of
(Krishnan et al., 2019), Improved SambaMahsuri (Madhavi et al., genotype–environment (GxE) interaction, density of markers,
2016), and Swarna-Sub1 (Neeraja et al., 2007) in rice, HUW510 in and genetic relationship between training population and
wheat (Vasistha et al., 2017), and HHB67-Improved in pearl breeding population or selection candidates, which may
millet (Rai et al., 2008). C214 in chickpea (Varshney et al., 2014a), influence the genomic prediction’s accuracy (De Roos et al.,
JTN5503 and DS880 in soybean (Arelli et al., 2006, 2009), and 2009; Lorenzana and Bernardo, 2009; Luan et al., 2009;
JL24 and TAG24 in groundnut (Varshney et al., 2014b) have been Daetwyler et al., 2010; Clark et al., 2011; Howard et al., 2014).
derived using MAS. However, MAS is practically feasible only if Hence, successful implementation of GS in breeding programs
the trait of interest is associated with one or very few major genes, requires careful consideration of all these factors. Apart from
and it is impractical or irrelevant for quantitative traits these factors, there are certain limitations of genomic selection.
(i.e., polygenic traits that are governed by few hundreds of Changes in gene frequencies and epistatic interactions drastically
minor genes) (Bernardo, 2008), which most of the stress affect the estimates of GEBV. Most of the models used to estimate
tolerance–related traits are based on. To overcome this issue, a GEBV ignore the effect of epistasis which plays a prime role
Frontiers in Genetics | www.frontiersin.org 2 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
especially in cross pollinated plants (Heffner et al., 2009). The rate One major problem in linear models using several thousands
of declination of selection response is more in GS than pedigree of genome-wide markers is that the number of markers (p)
based selection, which can be minimized through the addition of exceeds the number of observations (n), i.e., genotype/
new markers to the model (Nakaya and Isobe, 2012). However, individuals/lines, and this creates the problem of over-
the cost of implementation of GS is more than that of the parameterization (large “p” and small “n” problem (p >> n)).
traditional breeding program. Using a subset of significant markers can be an alternative for
The choice of models is an important factor in implementing dealing with the large “p” and small “n” problem.Meuwissen et al.
GS, and several parametric and non-parametric genomic (2001) used a modification of the least-squares regression for GS.
prediction models are available for this purpose. One of the They performed least-squares regression analysis on each marker
most common and widely used parametric genomic selection separately with the following model:
model is the best linear unbiased prediction (BLUP). It is a
Y 
 Xjβj + ε
mixed model–based whole-genome regression approach that is
used to estimate the marker effects, and the same has been whereXj 
 jth column of the designmatrix of the markers and β
successfully applied to predict complex traits (Habier et al., j
= genetic effect of the jth marker.
2009, 2013; de los Campos et al., 2013). In general, it was Markers with significant effects are selected using the log
observed that the performance of parametric models found to be likelihood of this model, and those are further used for
efficient only for traits with additive genetic architectures. For estimation of breeding values. However, it has to be noted that
traits that are highly affected by epistatic or non-additive some key informationmay be lost by selection based on the subset
interactions, it becomes challenging to use parametric models of markers.
(Moore and Williams, 2009). Epistatic interactions play a key Hence, an efficient solution for the over-parameterization
role in explaining genetic variation for quantitative traits. problem in linear models is using ridge regression (RR), which
Hence, ignoring such type of information in the prediction is a penalized regression–based approach (Meuwissen et al.,
model might result in lower genomic prediction accuracies 2001). It also solves the problems of multicollinearity at the
(Cooper et al., 2002). Due to these factors, it is not always same time (i.e., correlated predictors, e.g., SNP, or markers).
advisable to practice simple linear or parametric models. RR shrinks the coefficients of correlated predictors equally
Gianola et al. (2006) first used non-parametric and toward zero and solves the regression problem using ℓ2
semiparametric methods for modeling the complex genetic penalized least squares. Here, the goal is to derive an estimator
architecture. Subsequently, several statistical methods were of parameter β with a smaller variance than the least-squares
implemented to model both additive and epistatic effects for estimator. Similar to RR, the least absolute shrinkage and
genomic selection (Xu, 2007; Cai et al., 2011). For a detailed selection operator (LASSO) (Tibshirani, 1996; Usai et al.,
comparison of various parametric, non-parametric and 2009) is another variant of penalized regression, which uses
semiparametric methods in different settings of population the ℓ1 penalized least-squares criterion to obtain a sparse
size and trait heritability, one can refer to Howard et al. solution. However, sometimes LASSO may not work well with
(2014) and Budhlakoti et al. (2020c). Recently, some highly correlated predictors (e.g., SNPs in high linkage
semiparametric (Legarra and Reverter, 2018) and advanced disequilibrium) (Ogutu et al., 2012). The elastic net (ENET) is
approaches (Tanaka, 2018; Budhlakoti et al., 2020a, 2020b; an extension of the LASSO that is robust to extreme correlations
Majumdar et al., 2020; Sehgal et al., 2020; Tanaka, 2020; among the predictors (Friedman et al., 2010), and it is a
Mishra et al., 2021) have also been proposed and compromise between ℓ1 penalty (LASSO) and ℓ2 penalty (RR)
implemented in context to genomic selection. In the next (Zou and Hastie, 2005).
section, few most commonly used methods for genomic The RRmodel considers that each marker contributes to equal
selection studies have been discussed. variance, which is not the case for all traits. Therefore, the
variance of the markers based on the trait’s genetic
architecture has to be modeled. For this purpose, several
STATISTICAL MODEL FOR GENOMIC Bayesian models have been proposed where it is assumed that
SELECTION there is some prior distribution of marker effects. Furthermore,
inferences about model parameters are obtained on the basis of
The process of selecting the suitable individuals in GS starts with a posterior distributions of marker effects. There are several
simple linear model sometimes also called least-squares variants of Bayesian models for genomic prediction such as
regression or ordinary least-squares regression (OLS): Bayes A, Bayes B, Bayes Cπ, and Bayes Dπ (Meuwissen et al.,
Y 
 1nµ +Xβ + ε 2001; Habier et al., 2011) and other derivatives, e.g., Bayesian
LASSO and Bayesian ridge regression (BRR). Besides the marker-
where Y 
 n × 1 vectors of observations, µ is the mean, β 
 p × 1 based models, the best linear unbiased prediction (BLUP)
vectors of marker effects, ε 
 n × 1 vectors of random residual (Henderson et al., 1959) is one of the most commonly used
effects, X = design matrix of order n × p (where each row genomic prediction methods. There are many variants of BLUP
represents the genotype/individuals/lines (n) and each column available for this purpose, e.g., genomic BLUP (GBLUP), single-
corresponds to the marker (p)), and ε N(0, σ2e). step GBLUP (ssGBLUP), ridge regression BLUP (RRBLUP), and
˜
Frontiers in Genetics | www.frontiersin.org 3 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
FIGURE 2 | Overall summary of the most commonly used models in genomic selection.
GBLUP with linear ridge kernel regression (rrGBLUP), of which mixed model approach (Jia and Jannink, 2012; Klápště et al.,
GBLUP is very frequently used. The GBLUP uses the genomic 2020), Bayesian multi-trait model (Jia and Jannink, 2012; Cheng
relationships calculated using markers instead of the et al., 2018), MRCE (multivariate regression with covariance
conventional BLUP which uses the pedigree relationships to estimation) (Rothman et al., 2010), and cGGM (conditional
obtain the GEBV of the lines or individuals (Meuwissen et al., Gaussian graphical model) (Chiquet et al., 2017). Jia and
2001). Jannink (2012) presented three multivariate linear models
The genomic prediction models discussed so far perform well (i.e., GBLUP, Bayes A, and Bayes Cπ) and compared them to
for traits with additive genetic architecture, but their performance univariate models, and a detailed comparison of various STGS-
becomes very poor in case of epistatic genetic architectures. and MTGS-based methods has also been studied by Budhlakoti
Hence, Gianola et al. (2006) first used non-parametric and et al. (2019c). A brief structure of different STGS- and MTGS-
semiparametric methods for modeling the complex genetic based methods used in GS studies is given in Figure 2.
architecture. Subsequently, several statistical methods were
implemented to model both additive and epistatic effects for
genomic selection (Xu, 2007; Cai et al., 2011; Legarra and GS: IMPLICATIONS IN CROP
Reverter, 2018). There are several non-parametric methods IMPROVEMENT
that have been studied in relation to genomic selection, e.g.,
NW (Nadaraya–Watson) estimator (Gianola et al., 2006), RKHS GS in Cereals
(reproductive kernel Hilbert space) (Gianola et al., 2006), SVM Cereals are an important part of our daily diet as they contribute
(support vector machine) (Maenhout et al., 2007; Long et al., about 50% of the total dietary energy supply (WHO/FAO, 2003).
2011), ANN (artificial neural network) (Gianola et al., 2011), and Wheat, rice, maize, and barley are the major cereal crops, which
RF (random forest) (Holliday et al., 2012), among them SVM, are being grown on arable land all over the world amounting to a
NN, and RF are based on the machine learning approach. total of 2,817 million tonnes of production (FAO). Production of
Methods discussed earlier in this section are based on genomic these crops is being challenged by calamities created by a change
information where information is available for a single trait, in climatic pattern (Reynolds, 2010), and over that, it is being
i.e., single-trait genomic selection (STGS). As the performance complicated by the rising demand of increasing population
of STGS-based methods may be affected significantly in case of (Tester and Langridge, 2010; Furbank and Tester, 2011). To
pleiotropy, i.e., one gene linked to multiple traits, a mutation in a meet the challenges, the production system has to be efficient
pleiotropic gene may have an effect on several traits and sustainable with lower pressure on the ecosystem. High-
simultaneously. It was observed that low heritability traits can yielding, resource-efficient crop varieties are an integral
borrow information from correlated traits and consequently component of such production systems which can address the
achieve higher prediction accuracy. However, STGS-based challenges. But the development of such variety is a painstaking
methods consider the information of each trait independently. endeavor as most of the crop productivity traits are under the
Hence, we may lose crucial information which may ultimately control of a complex genetic system (most genes are of minor
result in poor genomic prediction accuracy. Nowadays, as we are effect) with the complication of low heritability and high order of
receiving data on multiple traits, so multi-trait genomic selection epitasis (Mackay, 2001). Though conventional selection methods
(MTGS)-based methods may provide more accurate GEBV and have resulted in a number of varieties but the genetic gain per unit
subsequently a higher prediction accuracy. Several MTGS-based time is not as much rewarding as GS, it provides an opportunity
methods have been studied in relation to GS, e.g., multivariate to hasten the cycle of selection (Bernardo and Yu, 2007; Lorenz
Frontiers in Genetics | www.frontiersin.org 4 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
et al., 2011). The potential of GS can be assessed from the fact that statistical models, marker platforms, types of populations used,
it has the ability to select high breeding value individuals rapidly and the prediction accuracies of statistical models are listed in
from early-generation populations without the need of extensive Table 1.
phenotyping. This has been shown effectively in cereal crops in
the recent past. Wheat, rice, maize, and barley are the first Biotic Stress Tolerance
candidate crops where the effectiveness of GS has been With the change in weather patterns, emergence/resurgence of
studied. GS in these crops leads to the identification of new races and biotypes of pathogens and insects is being reported
different models which were able to efficiently predict the globally (Juarez et al., 2013; Váry et al., 2015; Fones et al., 2020).
performance of traits under question and filter out the Hence, identification of resistance genes in the germplasm and
important breeding material. In the following section, the role their incorporation into the breeding program are required to
of GS in cereal crops has been discussed. develop biotic stress–tolerant varieties. MAS has proved to be
efficient in breeding for qualitative resistance, but for quantitative
Grain Yield and Related Traits’ Improvement resistance which is governed by many genes with smaller effects,
Grain yield is a major trait which is affected directly or indirectly MAS has not been so effective. GS has proved its role in
by other traits including thousand grain weight, number of tillers improving tolerance against biotic stresses in cereals which are
bearing panicle, number of grains per panicle, number of filled quantitatively controlled, though it has been applied to a very
grains per panicle etc. Genomic prediction for these traits limited extent. Most of the studies on the utility of GS for biotic
utilizing different types of training populations and models stress tolerance have been reported from wheat, for a wide array
have been evaluated. The variations in the accuracies of of diseases including three types of rusts, Fusarium head blight,
genomic prediction have been attributed to the heritability of septoria tritici blotch, powdery mildew, tan spot, and
the trait, training population, and models used. The genomic Stagonospora nodorum blotch. The genomic prediction
prediction accuracy for a very complex and physiological accuracies for these diseases ranged from 0.14 to 0.85
trait–like distribution of weight to the individual grain in the (Rutkoski et al., 2012; Daetwyler et al., 2014; Mirdita et al.,
panicle in rice (Yabe et al., 2018) ranged from 0.28 to 0.78 for 2015; Juliana et al., 2017; Sarinelli et al., 2019). In rice, GS has
grain yield in maize (Rio et al., 2019). For the improvement of been utilized to identify blast-tolerant lines (Huang et al., 2019).
accuracy, the role of training population also has a significant In maize, GS has been successfully utilized to select lines from
effect, and it has been reported that prediction based on the natural populations for tolerance to Stenocarpella maydis causing
training set developed using North Carolina mating design II ear rot (dos Santos et al., 2016) and from biparental populations
(0.60) was found at par with that of full diallel matings (0.58) and for superior yield under heavy infestation of Striga (Badu-Apraku
superior to that of test cross (0.10) (Fristche-Neto et al., 2018). et al., 2019). In case of barley, markers and prediction models
Similarly, better prediction accuracies for grain yield were were utilized for Fusarium head blight severity, and the
observed in recombinant inbred lines and doubled haploid prediction accuracy was quite higher, i.e., 0.72, than that of
populations compared to natural populations (Liu et al., 2018). conventional phenotyping (Lorenz et al., 2012; Sallam and
The accuracy of GS for grain yield is also highly influenced by the Smith, 2016).
size of training populations and genetic relationships between the
training and breeding populations (Lozada et al., 2019; Lozada Abiotic Stress Tolerance
and Carter, 2020). Longin et al. (2014) reported that GS followed The occurrence of drought, high-temperature stress during crop
by one cycle of phenotypic selection has been reported to facilitate growth stages, flood, etc., is at surge due to climate change,
identification of superior parental lines with better combining causing significant crop losses (Qin et al., 2011). With the 1°C
ability and high annual genetic gain for grain yield in wheat than increase in global temperature, yield reduction has been predicted
simple phenotypic selection. However scheme had not up to 6.4% in wheat (Liu et al., 2016). The sustainable and
considered the cost and time involved in production and economic options under such situations to cover the losses are
nursery screening of these lines, and thus, additional schemes changing cropping patterns or developing abiotic stress–tolerant
like GSrapid have been proposed which have better selection gain varieties. Identification of tolerant genotypes from the germplasm
and have been recommended for utilization in a hybrid breeding and their utilization in the breeding program become a prime
program of different cereal crops (Marulanda et al., 2016). GS requirement for development of such varieties (Baenziger, 2016).
could also be potentially used in the prediction of the The major issue in breeding for abiotic stress tolerance is their
performance of a large number of hybrid combinations complex inheritance, low heritability, and high environmental
(VanRaden, 2008; Crossa et al., 2017). The earlier GS studies effect on them (Bernardo, 2008).
on cereals started with wheat where the DArT marker system was Conventional breeding methods for abiotic stresses suffer
used (Crossa et al., 2010, 2011; Heffner et al., 2011; Burgueño from limitations of accuracy and reproducibility. Though
et al., 2012; Pérez-Rodríguez et al., 2012). However, later, other molecular markers have been utilized to identify and transfer
genome-wide SNP platforms became the routine marker in yield QTLs under abiotic stress conditions (Ribaut and Ragot,
genomic selection owing to their own advantages (Poland 2007; Almeida et al., 2013), but it may not be effective as QTL
et al., 2012; Zhao et al., 2012). Detailed information on GS from limited genetic resources explain little variation for grain
studies for grain yield and related traits in major cereals, yield under stress and are also highly influenced by the genetic
pulses, oilseeds, and horticultural crops with the details of background (Semagn et al., 2013) as well as the environment and
Frontiers in Genetics | www.frontiersin.org 5 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
Frontiers in Genetics | www.frontiersin.org 6 February 2022 | Volume 13 | Article 832153
TABLE 1 | Genomic prediction for grain yield and related traits in different crops (i.e. Cereals, Pulses, Oilseeds and Horticultural crops).
Crop Model Genotyping Techniques Population type Trait Prediction accuracy (PA) Reference
A. Cereals
i) Maize GBLUP Taqman (ABI 2002) F1 from half diallel and test crosses Grain yield (GY) 0.58 Zhao et al. (2012)
GBLUP Affymetrix® F1 from test crosses (TC), North Carolina design II GY 0.10 (TC) Fristche-Neto et al.
(NCII), and full diallel (FD) 0.58(NCII) (2018)
0.60(FD)
RRBLUP 55 K SNP array Natural population (NP), recombinant inbred line (RIL), GY RIL&DH (0.41) > F2:3 (0.36) Liu et al. (2018)
double haploid (DH), and F2:3 > NP(0.40)
GBLUP 100 kernel weight F2:3 (0.77) > RIL&DH (0.65)
> NP(0.48)
Bayes A, Bayes B, Bayes C, LASSO, and RKHS ()
GBLUP and multigroup GBLUP Genotyping by TC GY 0.78 Rio et al. (2019)
sequencing (GBS)
50 K Illumina® Yield index (YI) 0.73
600 K and Affymetrix® Axiom
RRBLUP and BSSV (Bayesian stochastic search DArTSeq™ and Illumina Inbred lines Ear rot Proportion of rotten 0.87 dos Santos et al. (2016)
variable) HiSeq2000 kernel
Ear rot incidence 0.24–0.56
BLUP Kompetitive Allele Specific Inbred lines and test cross progenies Striga resistance 0.58 Badu-Apraku et al.
PCR (KASP) Drought tolerance 0.42–0.65 (2019)
GBLUP GBS Breeding lines Drought tolerance 0.37–0.38 Beyene et al. (2015)
RRBLUP and GBLUP KASP Inbred lines and half diallel population Water-logging tolerance 0.53–0.84 Das et al. (2020)
BLUP KASP Asian and African inbred lines Drought GY 0.71–0.75 Vivek et al. (2017)
tolerance Anthesis–silking 0.35–0.43
interval (ASI)
RRBLUP Infinium Maize SNP50 Bead Subtropical maize lines Drought ASI 0.93 Shikha et al. (2017)
Bayes A, Bayes B, and LASSO Chip tolerance 100 kernel weight 0.92
ii) Wheat RRBLUP, RKHS, and Bayesian LASSO Diversity Arrays Technology Advanced breeding and germplasm lines GY 0.49–0.61 Crossa et al. (2010)
(DArT)
Bayesian LASSO and RKHS DArT Breeding lines GY 0.43–0.79 Crossa et al. (2011)
Bayes A, Bayes B, Bayes C, and RRBLUP DArT Breeding lines GY 0.48 Heffner et al. (2011)
Bayesian LASSO DArT Breeding lines GY 0.5–0.6 Burgueño et al. (2012)
RRBLUP, Bayes A, Bayes B, Bayes C, LASSO, NN, and DArT Breeding lines GY 0.6–0.7 Pérez-Rodríguez et al.
RKHS (2012)
GBLUP DArT Breeding lines GY 0.2–0.4 Poland et al. (2012)
RRBLUP, Bayes A, Bayes B, and Bayes C 9 K Illumina® Infinium F1s GY 0.3–0.6 Zhao et al. (2013)
RRBLUP 9 K Illumina® and 90 K iSelect Red winter wheat breeding lines GY 0.14–0.43 Lozada et al. (2019)
GBLUP DArT and KASP F4:6 population GY 0.75 Michel et al. (2019)
GBLUP GBS Breeding lines GY 0.42–0.56 Juliana et al. (2019)
GBLUP GBS Breeding population GY 0.12–0.34 Sun et al. (2019)
GBLUP and IBCF:MTME (item-based collaborative Illumina® 90 K Winter wheat lines GY -0.21 to 0.42 Lozada and Carter
filtering: multi-trait multi-environment) (2020)
GBLUP and BRR Infinium iSelect 9 K Germplasm Leaf rust resistance (LRR) 0.35 Daetwyler et al. (2014)
Stem rust resistance (SRR) 0.27
Yellow rust resistance (YRR) 0.44
RR DArT Breeding lines Fusarium head blight (FHB) resistance 0.006–0.463 Rutkoski et al. (2012)
RKHS 0.118–0.575
RF Deoxynivalenol (DON) resistance
Bayesian LASSO and multiple linear regression
RRBLUP Illumina Infinium 9 K and 90 K Winter wheat breeding lines FHB resistance 0.6 Mirdita et al. (2015)
Bayes Cπ and RKHS Septoria leaf blotch resistance 0.5
RRBLUP GBS Winter wheat breeding lines Powdery mildew resistance 0.60 Sarinelli et al. (2019)
GY 0.64
Test weight 0.71
RKHS and GBLUP GBS Lines from International Bread Wheat Screening LRR Seedling 0.31–0.74 Juliana et al. (2017)
Nursery Adult 0.12–0.56
YRR Seedling 0.70–0.78
Adult 0.34–0.71
SRR 0.31–0.65
(Continued on following page)
Budhlakoti et al. A Comprehensive Review on Genomic Selection
Frontiers in Genetics | www.frontiersin.org 7 February 2022 | Volume 13 | Article 832153
TABLE 1 | (Continued) Genomic prediction for grain yield and related traits in different crops (i.e. Cereals, Pulses, Oilseeds and Horticultural crops).
Crop Model Genotyping Techniques Population type Trait Prediction accuracy (PA) Reference
iii) Rice Bayesian LASSO DArT Inter-related synthetic population GY 0.309 Grenier et al. (2015)
Panicle weight 0.327
RRBLUP GBS Tropical rice breeding lines GY 0.31 Spindel et al. (2015)
GBLUP Illumina HiSeq 2000 128 Japanese rice varieties Field grain 0.30 Yabe et al. (2018)
Field grain weight 0.28
Illumina HiSeq 4000 and Variance of field grain 0.53
HiSeqX
GBLUP, SVM, LASSO, and PLS GBS North Carolina design II population GY ~0.5 Xu et al. (2018)
Thousand grain weight (TGW) ~0.28
GBLUP Illumina® HiSeq 2000 Hybrid population GY 0.54 Cui et al. (2020)
Grain length 0.92
GBLUP, RKHS, and Bayes B GBS Breeding lines Panicle weight 0.30 Hassen et al. (2018)
Nitrogen balance index 0.21
GBLUP SNP Breeding lines GY 0.39 Wang et al. (2018)
TGW 0.88
RRBLUP and GBLUP GBS Rice population Blast resistance 0.17–0.73 Huang et al. (2019)
GBLUP and RKHS 962 K Core SNP dataset Germplasm Drought tolerance 0.226–0.809 Bhandari et al. (2019)
iv) Barley RRBLUP Illumina GoldenGate Breeding lines GY 0.57 Sallam et al. (2015)
DON 0.72
FHB 0.74
GBLUP and RKHS GBS Breeding lines Thousand kernel weight (TKW) 0.67 Abed et al. (2018)
GBLUP Illumina Breeding lines GY 0.362 Tiede and Smith (2018)
DON resistance 0.367
B. Pulses
i) Lentil RRBLUP Exome capture Lentil diversity panel, RIL Maturity duration 0.58–0.84 Haile et al. (2020)
GBLUP
Bayes A
Bayes B
Bayes Cπ
Bayesian LASSO
BRR and RKHS
ii) Common GBLUP GBS RIL, multi-parent advanced generation inter-cross Cooking time 0.22–0.55 Diaz et al. (2021)
bean Bayes A (MAGIC), germplasm
Bayes B
Bayes C
Bayesian LASSO and BRR
RKHS GBS Breeding lines Root rot Fusarium 0.52 Diaz et al. (2021)
resistance Pythium 0.72–0.79
iii) Chickpea RRBLUP Whole-genome re-sequencing Breeding lines Drought tolerance 0.56–0.61 Li et al. (2018)
Bayesian LASSO and BRR (WGRS)
C. Oilseeds
i) Groundnut Bayesian generalized linear regression Affymetrix GeneTitan® Breeding lines Yield 0.49–0.60 Pandey et al. (2020)
Protein 0.41–0.46
Rust resistance 0.74–0.75
Late leaf spot resistance 0.57–0.65
ii) Brassica RRBLUP Infinium Array 60 K Test cross F1s Seed yield 0.45 Jan et al., 2016
napus Oil content 0.81
Lodging resistance 0.39
GBLUP Transcriptome GBSt assay Spring canola lines Seed yield 0.69 Fikere et al. (2020)
Oil content 0.64
GBLUP Illumina Infinium 60 K Double haploid population Seed yield 0.27–0.55 Xiong et al. (2020)
LASSO
iii) Sunflower and multi-kernel BLUP GBS F1s from factorial mating design Oil content 0.783 Mangin et al. (2017)
iv) Soybean RRBLUP iSelect Bead Chip RILs from interspecific cross Yield 0.68 Beche et al. (2021)
Oil content 0.76
Bayes B and Bayesian LASSO BARCSoySNP6K Protein content 0.76
RRBLUP iSelect Bead Chip Breeding lines Oil content 0.30 Stewart-Brown et al.
BARCSoySNP6K Protein content 0.55 (2019)
(Continued on following page)
Budhlakoti et al. A Comprehensive Review on Genomic Selection
there interactions. GS is superior to MAS, and the prediction
efficiency is also higher for abiotic stress tolerance (Cerrudo et al.,
2018). The usefulness of GS has been shown in wheat, maize, and
rice for drought and heat tolerance.
Beyene et al. (2015) have reported a gain of 0.086 t/ha for grain
yield, following the rapid cycling GS strategy in eight biparental
populations of maize under drought conditions, and a final gain
of 0.176 t/ha after three cycles of selection. This increased the
genetic gain as the time required for selection was reduced
significantly as compared to that of the conventional breeding
scheme, where it was three times higher with phenotypic
selection. Similarly, Das et al. (2020) reported a genetic gain of
0.110 and 0.135 t/ha/yr for grain yield under drought and 0.038
and 0.113 t/ha/yr under water logging in two maize populations,
viz., Maize Yellow Synthetic 1 and Maize Yellow Synthetic 2,
respectively, following rapid cycling genomic selection. Vivek
et al. (2017) compared the performances of second cycle selection
through phenotypic and rapid cycle genomic selection and found
10–20% superiority using the latter. Genomic prediction
accuracies using multi-environment models for drought stress
tolerance were higher than those using single-environment
models in rice and wheat (Sukumaran et al., 2018; Bhandari
et al., 2019). Prediction accuracies were higher for heat and
drought stress in case of wheat when secondary traits
contributing to yield were considered under stress rather than
yield per se using genomic prediction (Rutkoski et al., 2016).
Comparative analysis among different models leads to the
conclusion that multi-trait models are superior when selection
is carried out in severe drought conditions, while the random
regression model was better than the repeatability model and
multi-trait model under normal drought conditions and also use
of secondary high-throughput traits in genomic prediction
improved accuracies by ~70% (Sun et al., 2017).
Quality Improvement
Quality traits have varied genetic architectures, some being
controlled oligogenically like grain color, while others are
polygenic in nature, viz., grain size and protein content
(Battenfield et al., 2016). GS has been carried out in wheat
extensively for quality-related traits, viz., milling and flour
quality, and when prediction accuracies were compared in
biparental and multi-family populations, it was concluded that
the prediction accuracies in multi-family populations were better
(Heffner et al., 2011).
Protein content is known to be negatively correlated with yield
due to physiological compensation (Lam et al., 1996). Michel et al.
(2019) employed multi-trait genomic selection for grain yield,
protein content, and dough rheological traits for efficient
selection with optimized yield and protein content with better
quality. The prediction accuracy for the quality traits depends on
variability in the germplasm, the relationship among training and
prediction populations, etc. (Crossa et al., 2014; Zhao et al., 2015).
Joukhadar et al. (2021) used Bayesian regression and BRR for
rapid improvement of grain yield as well as mineral content to
biofortify wheat and reported Bayesian regression was better in
predicting mineral content with an accuracy of 0.55. In rice, grain
length and width are important quality parameters, and the
Frontiers in Genetics | www.frontiersin.org 8 February 2022 | Volume 13 | Article 832153
TABLE 1 | (Continued) Genomic prediction for grain yield and related traits in different crops (i.e. Cereals, Pulses, Oilseeds and Horticultural crops).
Crop Model Genotyping Techniques Population type Trait Prediction accuracy (PA) Reference
D. Horticultural crops
i) Apple RRBLUP HiScan Illumina and Infinium Germplasm and biparental families Firmness 0.81 Roth et al. (2020)
Array 20 K
RRBLUP and Bayesian LASSO Infinium® II 8 K F1 from factorial mating design Fruit firmness 0.83 Kumar et al. (2012)
Soluble solids 0.89
ii) Citrus GBLUP GBS F1 from different parental lines Fruit weight 0.650 Imai et al. (2019)
Sugar content 0.519
Acid content 0.666
GBLUP Illumina HiSeq 2000 Varieties and their full sib families Fruit weight distribution 0.89 Minamikawa et al. (2017)
iii) Apricot GBLUP GBS F1 pseudo-testcross population Glucose content 0.31 Nsibi et al. (2020)
RRBLUP 0.78
Bayes A, Bayes B, Bayes C, Bayesian LASSO, and BRR Ethylene content
iv) Pear GBLUP GBS Full sib families Crispness 0.32 Kumar et al. (2019)
Sweetness 0.62
v) Capsicum GBLUP, Bayesian LASSO, Bayes B, Bayes C, and RKHS GBS Core collection and RIL population Fruit length 0.32 Hong et al. (2020)
Fruit width 0.50
Fruit shape 0.34
Fruit weight 0.48
vi) Tomato RRBLUP Infinium Assay Germplasm Fruit weight 0.814 Duangjit et al. (2016)
Firmness 0.614
Soluble solids 0.714
Sugar content 0.649
Acidity 0.619
Biochemical profile 0.126–0.705
Budhlakoti et al. A Comprehensive Review on Genomic Selection
prediction accuracy for these traits ranged from 0.35 to 0.45 and their predictive accuracy. When each allele is distributed equally
0.5 to 0.7, respectively, in 110 Japanese rice cultivars employing in the population, the predictive accuracy for both the alleles is
various GS models (Onogi et al., 2015). In barley, the prediction the same. In such cases, it is obvious that the less frequent allele’s
for quality traits like malting quality (prediction accuracy: prediction is biased downward. Contiguous breeding programs
0.4–0.8) has shown the prospects of GS for screening large are very common where new cross combinations are added each
populations without the need of cost-intensive phenotyping year. In such cases, using nested association mapping (NAM)
(Schmidt et al., 2016). population is better in terms of prediction accuracy (for yield
0.68 and oil and for protein content 0.76) than biparental
GS in Oilseeds population, showing the potential of NAM where
Oilseeds are a source of livelihood to the smallholder farmers in connectedness is there among the population on the basis of
developing countries of Asia and Africa. The yield potential is still the common parent (Beche et al., 2021). Similarly, Stewart-
to be realized by bridging the yield gap via inducing tolerance to Brown et al. (2019) have reported that, for better predictions in
biotic and abiotic stresses and improvement in quality (Janila soybean, it is important to have good relatedness among
et al., 2016). Different traits related to biotic and abiotic stresses training and breeding populations. They have observed that
have been mapped, but most of them are qualitative in nature, the size of the training population has a larger effect on the
and the report of GS is limited in such potential crops. Oil quality prediction accuracy, compared to the marker density, but
and yield traits are influenced by the environment and GxE increasing the training population sizes beyond a limit had a
interactions (Patil et al., 2020). Hence, it is important to use the diminishing return on the prediction accuracy. Hu et al. (2011)
appropriate GS models to account for the GxE effects for accurate applied GS for biological process, i.e., embryogenesis capacity in
selection. Pandey et al. (2020) employed GS in groundnut with soybean, and reported a good prediction accuracy (0.78).
different models and validation schemes to account for GxE
interaction effects. The model having genomic information GS in Pulses
generated from the SNP (G), genotypic effect of the line (L), In lentil, Haile et al. (2020) showed that if large-effect QTLs were
environment effect (E), and their interactions (LxE and GxE) had present in the population, multi-trait–based Bayes B is the best
better mean accuracy (0.58) for all the traits compared to other GS model, while single-trait GS (STGS) is suitable in their
models. Jan et al. (2016) employed the RRBLUP model for GS in absence. They also reported that, for low heritable traits with
Brassica using 950 cross combinations derived from utilizing 475 GxE interactions, MTGS improves predictability. Considering
lines and two testers, for the improvement of oil-specific traits, quality traits in Phaseolus, i.e., cooking time for screening of fast
and the accuracy for oil content and oil yield was 0.81 and 0.75, culinary genotypes, Diaz et al. (2021) evaluated GS using different
respectively. Hence, they concluded that the GS model is helpful populations (RIL, MAGIC, Andean, and Mesoamerican breeding
in pre-selecting superior cross combinations before extensive lines). The trait was highly heritable (0.64–0.89), and genomic
field evaluation over location and years saving resources. prediction accuracies for cooking time using MAGIC population
Fikere et al. (2020) employed GS for 22 traits related to yield, were promising and high (0.55) compared to those of
disease resistance, and quality in B. napus and reported prediction Mesoamerican genotypes (0.22).
accuracy was highest for yield (0.69) followed by oil content Under the circumstance of less connectedness in the training
(0.64) using GBLUP. They also evaluated genomic prediction for and prediction populations, markers generated using the whole
compositional fatty acid estimated under rainfed and irrigated genome re-sequencing (WGRS) platform increase the
conditions and concluded that the prediction accuracies for these prediction accuracy; however, Li et al. (2018) proposed first
traits were lower under non-irrigated conditions. Xiong et al. identifying causal variants and then utilizing them into the
(2020) employed various prediction models, viz., LASSO, prediction. The prediction accuracy was 0.148–0.186 for yield
GBLUP, OLS, and OLS post-LASSO, for different traits in B. under drought when using all the SNP from WGRS, but when
napus and reported the two-stage method OLS post-LASSO to be filtered yield-related causal SNPs were employed, it was
the most accurate (0.90 and 0.55 for oil content and single plant observed that prediction accuracy significantly improved
yield, respectively) with the provision of incorporating GxE (0.56–0.61). Diaz et al. (2021) employed GS for root rot
interactions. For oil content in sunflower which is highly resistance and reported high prediction accuracies (0.7–0.8)
heritable and additive in nature, Mangin et al. (2017) reported for both rots (Pythium and Fusarium) in Phaseolus and
that accuracy based on general combining ability (GCA) and GS proposed it to be promising for improving quantitative
were on par, and in case if there is no knowledge about one of the tolerance.
parents of hybrid combination, GS excels the GCA-based
predictions. Similar inferences had been made by Reif et al. GS in Horticultural Crops
(2013) for the prediction of hybrid performance in sunflower. Fruit and vegetables are indispensable in achieving nutritional
From a cross between cultivated and wild progenitors of security. However, the problem associated with their breeding,
soybean (G. max X G. sojae), Beche et al. (2021) reported that especially of fruits, has its own limitations, viz., long juvenile
the yield-related alleles were associated with the cultivated elite phase and highly heterozygous nature. Therefore, genetic gain is
line, but the protein content alleles were from the wild not much as per the Lush equation. In such crops, GS can be a
progenitor. The difference in the distribution of trait- perfect tool where prediction of performance for quality- and
contributing alleles in such crosses has a greater impact on yield-related traits which are under a complex genetic system can
Frontiers in Genetics | www.frontiersin.org 9 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
be utilized to improve selection accuracy and efficiency in GMStool
developing varieties. The success of GS in annual crops has It is a genome-wide association study (GWAS)-based tool for
led the horticultural crop breeder to utilize its potential in genomic prediction using genome-wide marker data. It searches
perennial fruit as well as annual fruit and vegetable crops. for the optimum number of markers for prediction using
Roth et al. (2020) evaluated 537 genotypes in apple for fruit appropriate statistical and machine learning/deep
texture traits and performed GS and reported the accuracy up to learning–based models and chooses the best prediction model
0.81. It was suggested to have a large training population from (Jeong et al., 2020). Furthermore, it identifies SNP markers with
which a tailored training population with a priori genetic the lowest p-values (e.g., top 100 markers) in the GWAS and then
relatedness information and ample variation can be formed chooses the relevant markers set to be included in the final
and utilized to predict the performance of population under prediction model. GMStool is R-based and freely available
consideration. Kumar et al. (2012) have shown high prediction through the GitHub repository at https://github.com/
accuracy in apple for different quality traits utilizing a factorial JaeYoonKim72/GMStool. The whole process or its algorithm is
mating design (0.70–0.90). Imai et al. (2019) reported that basically divided into three steps: data preparation, marker
ssGBLUP predicts with higher accuracy (0.650, 0.519, and selection, and final prediction model. The detailed procedure
0.666) than GBLUP (0.642, 0.432, and 0.655) for quality traits of GMStool is discussed below.
in citrus, viz., fruit weight, sugar content, and acid content from Step 1: Input data are divided into training and test sets (user
population where some individuals are not genotyped using defined)
information from genotyped related individuals, hence Step 2: The training set is further divided into small datasets
reducing the cost at hand. for performing cross validation (i.e., k-folds, for example, five or
As fruits are perishable produce and the post-harvest ten folds) followed by marker selection in each group or fold. The
attribute of the fruits plays an important role in storability, process of marker selection is performed in each fold/group
attempts have been made to employ GS for such traits. In simultaneously.
apricot, Nsibi et al. (2020) reported prediction accuracy Step 3: The selected marker from each fold is integrated into
ranging from 0.31 to 0.78 for glucose content and ethylene the final marker set for updating the model. Appropriate
production. Minamikawa et al. (2017) compared different statistical and machine learning–based models are then used
models of GS for fruit weight distribution among two groups for genomic prediction.
of fruit sizes and reported that, among a large fruit size group,
rrGBLUP (0.89) was superior to GBLUP (0.74) and the same solGS
was in the case of a small fruit size group, i.e., rrGBLUP (0.32) It is an open-source tool based on the Linux operating system.
and GBLUP (0.30). Also, it was proposed to have breeding The workflow of the tool is broadly divided into two steps,
population or combined parental and breeding population as i.e., training of the prediction model and obtaining GEBV.
training population to have better accuracy than only having However, there are three approaches available for training the
parental as training population which was consistent for all the prediction model, i.e., trait-based approach, trial approach, and
quality-related traits. Kumar et al. (2019) employed GS in pear custom lists approach. Here, model input and output could be
for various fruit quality traits ranging from texture to taste and visualized graphically and can be interactively explored or
observed the prediction accuracy ranged from 0.32 to 0.62 downloaded. It is designed to store a large amount of
averaging to 0.42 and also suggested that training population genotypic, phenotypic, and experimental data. In the
should be multi-generational and evaluated rigorously over background, it basically uses two R-based packages, i.e., nlme
location and time, to have better prediction accuracy. (Pinheiro et al., 2017) for data preprocessing and rrBLUP
Various GS models have been evaluated for different fruit- (Endelman, 2011) for statistical modeling. solGS was earlier
related traits in capsicum and reported that RKHS had better used by the NEXTGEN Cassava project (http://nextgencassava.
accuracy ranging from 0.75 to 0.82 and positively correlated org) and implemented at the Cassavabase website (http://
with the number of markers (Hong et al., 2020). GS is also cassavabase.org/solgs).
performed to evaluate the accuracy of prediction of different
biochemical parameters important for fruit quality in tomato rrBLUP
which ranged from 0.13 to 0.70 for aspartate content and also It is an R package based on BLUP, which is a mixed linear model
for other traits, viz., fruit weight (0.81), firmness (0.61), soluble framework (Endelman, 2011). It is one of the most widely used
solids (0.71), sugar content (0.65), and acidity (0.62) (Duangjit packages for genomic prediction in animal and plant breeding.
et al., 2016). This package estimates the marker effects from training datasets
and ultimately estimates the GEBV for the selection candidates.
The mixed.solve function, a linear mixed model equation which
STATISTICAL TOOLS FOR IMPLEMENTING estimates marker effects and GEBV, is one of the most commonly
GENOMIC SELECTION used functions of this package. An additive relationship matrix of
individuals can be calculated using genotypic data for the
Several tools and packages have been developed for the evaluation estimation of GEBV using GBLUP. rrBLUP is an open-source
of genomic prediction and implementation of GS, some of which package and can be accessed at https://CRAN.R-project.org/
are discussed below. package=rrBLUP.
Frontiers in Genetics | www.frontiersin.org 10 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
BWGS com/perpdgo/lme4GS. It is an extension of the lme4 R
It is an integrated pipeline based on R and freely available at package, which is the standard package for fitting linear
https://CRAN.R-project.org/package=BWGS. The BWGS mixed models. lme4GS package is basically motivated from
(i.e., BreedWheat Genomic Selection) pipeline (Charmet existing R packages pedigreemm (Vazquez et al., 2010) and
et al., 2020) basically consists of three modules: i) missing lme4qtl (Ziyatdinov et al., 2018). lme4GS package can also be
data imputation, ii) dimension reduction, i.e., reducing the considered an extension of the rrBLUP (Endelman, 2011)
number of markers as it could enhance the speed of package. Further, lme4GS package can be used for fitting
computation on large datasets, and iii) estimation of GEBV. mixed models with covariance structures defined by the
It has a wide choice of totally 15 parametric and non- user, bandwidth selection, and genomic prediction.
parametric statistical models for estimation of GEBV for
selection candidates. It could be used for estimation of STGS
GEBV for a wide range of genetic architectures. This tool It is an R-based package developed for genomic predictions by
comprises mainly two functions: bwgs.cv and bwgs.predict. The estimating marker effects, and the same is further used for
former is used for missing value imputation, dimension calculation of genotypic merit of individuals, i.e., GEBV. GS
reduction, and cross validation, while the later is used for may be based on single-trait or multi-trait information. This
model calibration and estimation of GEBV for selection package performs genomic selection only for a single trait,
candidates. hence named STGS, i.e., single-trait genomic selection
(Budhlakoti et al., 2019a). STGS is a comprehensive
BGLR package which gives a single-step solution for genomic
This package is basically an extension of the BLR package (Perez selection based on most commonly used statistical methods
and Campos, 2014). It can be used to implement several Bayesian (i.e., RR, BLUP, LASSO, SVM, ANN, and RF). It is freely
models and also provides flexibility in terms of prior density available through the CRAN server at https://CRAN.R-
distribution. Here, the response to be considered could be project.org/package=STGS.
continuous or categorical (either binary or ordinal). It is freely
available in the public domain through the CRAN mirror at MTGS
https://CRAN.R-project.org/package=BGLR. It is an R-based package developed for genomic predictions by
estimating marker effects based on information available on
GenSel multiple traits. Currently, STGS methods could not utilize
The GenSel software program was developed and implemented additional information available when using multi-trait data.
under the BIGS (Bioinformatics to Implement Genomic The package MTGS performs genomic selection using multi-
Selection) project (Fernando and Garrick, 2009). It is used for trait information (Budhlakoti et al., 2019b). MTGS is a
estimation of molecular marker–based breeding values of animals comprehensive package which gives a single-step solution for
for the trait of interest. This can serve the purpose through the genomic selection using various MTGS-based methods (MRCE,
command line (MAC or Linux) interface or as a user-friendly MLASSO, i.e., multivariate LASSO, and KMLASSO,
tool. The jobs are submitted and assigned in the queue for i.e., kernelized multivariate LASSO). It is freely available
analysis. The software uses the Bayesian approach in the through the CRAN server at https://CRAN.R-project.org/
background for estimation of marker effects from the training package=MTGS.
data and further for estimation of GEBV for breeding candidates.
This software program can be accessed at https://github.com/
austin-putz/GenSel. FACTORS AFFECTING GENOMIC
GSelection PREDICTION: EFFECTS OF MARKER
This is an R-based package and is freely available at https:// DENSITY, POPULATION SIZE, TRAIT
CRAN.R-project.org/package=GSelection. The package ARCHITECTURE, AND HERITABILITY
comprises of a set of functions to select the important markers
and estimates the GEBV of selection candidates using an In general, increased marker density enhances the prediction
integrated model framework (Majumdar et al., 2019). The accuracy using most of the GS models such as BLUP, LASSO,
motivation behind this package is that not a single method machine learning–based, or deep learning–based methods.
performs best in case of all crop plants or animal breeding However, there may be a chance of slow convergence in
programs as they may have diverse genetic architectures, methods like Bayesian (Bayes A, Bayes B, Bayes Cπ, and
i.e., additive and non-additive genetic effects. This package has Bayes Dπ), where convergence in terms of MCMC
been developed by integrating the best performing model from (i.e., Markov chain Monte Carlo) iteration is required
each category of additive and non-additive genetic models. (Arruda et al., 2016; Zhang et al., 2017; Norman et al.,
2018; Zhang et al., 2019). Sometimes, low-density markers
lme4GS of a few hundreds to thousands also enable high prediction
lme4GS is an R-based package freely available and can be accuracies in breeding populations provided that there is a
accessed through the GitHub repository at https://github. strong LD among the markers; however, it may be trait specific
Frontiers in Genetics | www.frontiersin.org 11 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
and may vary with the architecture and heritability of studied accuracy as in the case of traits with moderate to high
traits (Lorenz et al., 2011; Werner et al., 2018). Also sometimes heritability. However, to achieve this goal, sometimes cost
keeping a very high density of markers may have economic may be a limiting factor, especially in developing countries.
constraints as incorporation of such aspects into evaluation of Moreover, it could be observed from the available literature
GS strategies is also necessary for a profitable and efficient GS. that even for low heritable and complex traits, the
Therefore, it is always difficult to give a benchmark for the performance of BLUP and its derivatives (e.g., GBLUP and
number of markers to be used in such genomic studies; RRBLUP), Bayesian methods (Bayes A, Bayes B, Bayes Cπ,
however, it is advisable to keep a moderate density, at least and Bayes Dπ), and RKHS seems to be robust as compared to
2000 SNPs, so that prediction accuracy could not be their counterparts (Crossa et al., 2010; Crossa et al., 2011;
significantly hampered (Abed et al., 2018). However, the Heffner et al., 2011; Poland et al., 2012; Zhao et al., 2013;
cost of genotyping can also be significantly reduced by Spindel et al., 2015; Crossa et al., 2017; Wang et al., 2018; Xu
increasing the level of multiplexing without paying any et al., 2018; Juliana et al., 2019; Lozada et al., 2019; Michel
penalty in terms of genomic prediction accuracy (e.g., et al., 2019), and at the same time, most of the models work
genotyping a single line by GBS (96-plex) can cost 3.75 and fine with highly heritable traits, although the most suitable
4.25 times less than using 9 K and 50 K arrays, respectively, in method is usually case-dependent. Sometimes missing
barley) (Abed et al., 2018). The position of SNPs and how they observations also poses a challenge in estimating GEBV.
are placed in genomic arrangements over the chromosome However, the issue of low heritable trait and missing
may have a key role, for example, SNPs located in the observation could be handled simultaneously, provided that
intergenic space are slightly better at capturing the data are available on multiple traits. In multiple traits, if we
underlying haplotype diversity related to SNPs located in have few traits with low heritability and at the same time we
the genic space as the intergenic space is a playground of have a good correlation with other highly heritable traits,
many important regulatory sequences, such as promoters and i.e., by using the appropriate MTGS-based model, we can
enhancers (Barrett et al., 2012; Abed et al., 2018). The use of borrow information from other traits. In such scenarios, by
high-quality SNP genotyping data (i.e., minor allele frequency using the MTGS model, we can estimate the GEBV more
(MAF)>0.1) could also be suggested to achieve a good precisely and accurately.
prediction accuracy.
Population size has a significant role in the prediction
accuracy whether it is conventional MAS or genomic
selection, especially training population. If the population CONCLUSION AND PROSPECTS
size or training population size is small, it is obvious that a
decrease in accuracy is expected because the model will poorly Genomic selection has shown its potential in plant and
estimate the marker effects and hence prediction accuracy. animal breeding research by increasing genetic gains in the
However, as an idea or estimate for the size of training last two decades. Revolution in terms of cheaper NGS
population as 2*Ne*L (where Ne is the effective population technologies has made it possible to sequence the crop and
size and L is the genome size in Morgan) and the number of animal genomes at a relatively low cost. It resulted in a
markers as 10*Ne*L to achieve a prediction accuracy of 0.9 and number of completely sequenced crop and animal genomes
reducing the size of the training population to 1*Ne*L results in with high-density SNP genotyping chips and their availability
a prediction accuracy of 0.7, provided that training population in the public domain, which may further boost the predictive
and breeding population are unrelated or both separated by ability of a GS model. Even after more than a decade in the
many generations (Meuwissen, 2009). However, for most of the field of genomic selection studies, still there is a lot of scope
cases, training population and breeding population are related, for improvement in this area. Methodological refinements
so high genomic prediction accuracy could be achieved with a (such as imputation of missing genotypic value,
training population size much smaller than that referred above implementation of GxE interaction, information on
(Meuwissen, 2009). epigenetic regulation, haplotypes, and including multi-trait
Apart from these factors, prediction accuracy can also be information into prediction models) will be definitely helpful
affected by trait heritability especially for lower heritability for a successful implementation of GS in plant and animal
(h2 < 0.4) (Hayes et al., 2009). Numerous studies up-to-date breeding programs. Consistent updation of the training set
showed that genomic selection accuracy is strongly influenced for GS is highly desirable by including the new markers in
by trait heritability, i.e., the fraction of the phenotypic each generation. Evaluation of the training populations
variance to the genetic variance of studied traits. Generally, should be done in controlled and well-managed conditions
it is assumed that the target trait with high heritability has as it significantly affects the performance of prediction
good prediction accuracies and vice versa. However, as most models. There is a need for a structured program in the
of the agricultural traits have low to moderate heritability, it field of genomic selection including human resource
poses a challenge to genomic selection studies, especially in development, advanced data recording methodologies, and
plants. However, low heritability traits would require a larger trait phenotyping in order to come out with fruitful
training population in order to attain the same prediction outcomes.
Frontiers in Genetics | www.frontiersin.org 12 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
AUTHOR CONTRIBUTIONS ACKNOWLEDGMENTS
NB, AK, AR, and DM contributed to conceptualization. NB, AK, We are grateful to the Government of India’s DBT project
KC, AK, AP, RK, and UK reviewed and edited the paper. PJ, DM, “Germplasm Genomics and Trait Discovery in Wheat” and
and SK contributed to the final editing and correction. All authors ICAR-CABin Scheme Network Project for the financial
contributed to themanuscript and approved the submitted version. support to carry out this study.
REFERENCES Budhlakoti, N., Mishra, D. C., Rai, A., and Chaturvedi, K. K. (2019a). Package
‘STGS’. 1–11. Available at: https://cran.r-project.org/web/packages/STGS/
STGS.pdf.
Abed, A., Pérez-Rodríguez, P., Crossa, J., and Belzile, F. (2018). When Less Can Be Budhlakoti, N., Mishra, D. C., and Rai, A. (2019b). Package ‘MTGS’. 1–6.
Better: How Can We Make Genomic Selection More Cost-Effective and Budhlakoti, N., Mishra, D. C., Rai, A., Lal, S. B., Chaturvedi, K. K., and Kumar, R. R.
Accurate in Barley? Theor. Appl. Genet. 131, 1873–1890. doi:10.1007/ (2019c). A Comparative Study of Single-Trait and Multi-Trait Genomic
S00122-018-3120-8 Selection. J. Comput. Biol. 26, 1100–1112. doi:10.1089/CMB.2019.0032
Almeida, G. D., Makumbi, D., Magorokosho, C., Nair, S., Borém, A., Ribaut, J.-M., Budhlakoti, N., Rai, A., and Mishra, D. C. (2020a). Effect of Influential Observation
et al. (2013). QTL Mapping in Three Tropical maize Populations Reveals a Set in Genomic Prediction Using LASSO Diagnostic. Indian J. Agric. Sci. 90,
of Constitutive and Adaptive Genomic Regions for Drought Tolerance. Theor. 1155–1159.
Appl. Genet. 126, 583–600. doi:10.1007/S00122-012-2003-7 Budhlakoti, N., Rai, A., Mishra, D. C., Jaggi, S., Kumar, M., and Rao, A. R. (2020c).
Arelli, P. R., Young, L. D., and Concibido, V. C. (2009). Inheritance of Resistance in Comparative Study of Different Non-parametric Genomic Selection Methods
Soybean PI 567516C to LY1 Nematode Population Infecting Cv. Hartwig. under Diverse Genetic Architecture. Ijgpb 80, 395–401. doi:10.31742/IJGPB.80.
Euphytica 165, 1–4. doi:10.1007/S10681-008-9760-Z 4.4
Arelli, P. R., Young, L. D., and Mengistu, A. (2006). Registration of High Yielding Budhlakoti, N., Rai, A., and Mishra, D. C. (2020b). Statistical Approach for
and Multiple Disease-Resistant Soybean Germplasm JTN-5503. Crop Sci. 46, Improving Genomic Prediction Accuracy through Efficient Diagnostic
2723–2724. doi:10.2135/cropsci2005.12.0471crg Measure of Influential Observation. Sci. Rep. 10, 1–11. doi:10.1038/s41598-
Arruda, M. P., Lipka, A. E., Brown, P. J., Krill, A. M., Thurber, C., Brown-Guedira, 020-65323-3
G., et al. (2016). Comparing Genomic Selection and Marker-Assisted Selection Burgueño, J., de los Campos, G., Weigel, K., and Crossa, J. (2012). Genomic
for Fusarium Head Blight Resistance in Wheat (Triticum aestivum L.). Mol. Prediction of Breeding Values when Modeling Genotype × Environment
Breed. 36, 84. doi:10.1007/s11032-016-0508-5 Interaction Using Pedigree and Dense Molecular Markers. Crop Sci. 52,
Badu-Apraku, B., Talabi, A. O., Fakorede, M. A. B., Fasanmade, Y., Gedil, M., 707–719. doi:10.2135/CROPSCI2011.06.0299
Magorokosho, C., et al. (2019). Yield Gains and Associated Changes in an Early Cai, X., Huang, A., and Xu, S. (2011). Fast Empirical Bayesian LASSO for Multiple
Yellow Bi-parental maize Population Following Genomic Selection for Striga Quantitative Trait Locus Mapping. BMC Bioinformatics 12, 211–213. doi:10.
Resistance and Drought Tolerance. BMC Plant Biol. 19, 129. doi:10.1186/ 1186/1471-2105-12-211/FIGURES/5
S12870-019-1740-Z Cerrudo, D., Cao, S., Yuan, Y., Martinez, C., Suarez, E. A., Babu, R., et al.
Baenziger, P. S. (2016). “Wheat Breeding and Genetics,” in Reference Module in (2018). Genomic Selection Outperforms Marker Assisted Selection for
Food Science. Editor C. Beddows (Amsterdam, Netherlands: Elsevier). doi:10. Grain Yield and Physiological Traits in a maize Doubled Haploid
1016/B978-0-08-100596-5.03001-8 Population across Water Treatments. Front. Plant Sci. 9, 366. doi:10.
Barrett, L. W., Fletcher, S., andWilton, S. D. (2012). Regulation of Eukaryotic Gene 3389/FPLS.2018.00366/BIBTEX
Expression by the Untranslated Gene Regions andOther Non-coding Elements. Charmet, G., Tran, L.-G., Auzanneau, J., Rincent, R., and Bouchet, S. (2020).
Cell. Mol. Life Sci. 69, 3613–3634. doi:10.1007/S00018-012-0990-9 BWGS: A R Package for Genomic Selection and its Application to a Wheat
Battenfield, S. D., Guzmán, C., Gaynor, R. C., Singh, R. P., Peña, R. J., Dreisigacker, Breeding Programme. PLOS ONE 15, e0222733. doi:10.1371/JOURNAL.
S., et al. (2016). Genomic Selection for Processing and End-Use Quality Traits PONE.0222733
in the CIMMYT Spring BreadWheat Breeding Program. Plant Genome 9, 1–12. Cheng, H., Kizilkaya, K., Zeng, J., Garrick, D., and Fernando, R. (2018). Genomic
doi:10.3835/PLANTGENOME2016.01.0005 Prediction from Multiple-Trait Bayesian Regression Methods Using Mixture
Beche, E., Gillman, J. D., Song, Q., Nelson, R., Beissinger, T., Decker, J., et al. Priors. Genetics 209, 89–103. doi:10.1534/GENETICS.118.300650/-/DC1
(2021). Genomic Prediction Using Training Population Design in Chiquet, J., Mary-Huard, T., Robin, S., and Robin, S. (2017). Structured
Interspecific Soybean Populations. Mol. Breed. 41, 1–15. doi:10.1007/ Regularization for Conditional Gaussian Graphical Models. Stat. Comput.
S11032-021-01203-6 27, 789–804. doi:10.1007/s11222-016-9654-1
Ben Hassen, M., Bartholomé, J., Valè, G., Cao, T.-V., and Ahmadi, N. (2018). Clark, S. A., Hickey, J. M., and Van Der Werf, J. H. (2011). Different Models of
Genomic Prediction Accounting for Genotype by Environment Interaction Genetic Variation and Their Effect on Genomic Evaluation. Genet. Sel. Evol. 43,
Offers an Effective Framework for Breeding Simultaneously for Adaptation to 18. doi:10.1186/1297-9686-43-18
an Abiotic Stress and Performance under normal Cropping Conditions in rice. Cooper, M., Podlich, D.W., Micallef, K. P., Smith, O. S., Jensen, N.M., Chapman, S.
G3 Genes, Genomes, Genet. 8, 2319–2332. doi:10.1534/g3.118.200098 C., et al. (2002). “Complexity, Quantitative Traits and Plant Breeding: a Role for
Bernardo, R. (2008). Molecular Markers and Selection for Complex Traits in Simulation Modelling in the Genetic Improvement of Crops,” in Quantitative
Plants: Learning from the Last 20 Years. Crop Sci. 48, 1649–1664. doi:10.2135/ Genetics, Genomics and Plant Breeding. Editor M. S. Kang (Wallingford, UK:
CROPSCI2008.03.0131 CAB International), 143–166. doi:10.1079/9780851996011.0143
Bernardo, R., and Yu, J. (2007). Prospects for Genomewide Selection for Crossa, J., Campos, G. d. l., Pérez, P., Gianola, D., Burgueño, J., Araus, J. L., et al.
Quantitative Traits in maize. Crop Sci. 47, 1082–1090. doi:10.2135/ (2010). Prediction of Genetic Values of Quantitative Traits in Plant Breeding
CROPSCI2006.11.0690 Using Pedigree and Molecular Markers. Genetics 186, 713–724. doi:10.1534/
Beyene, Y., Semagn, K., Mugo, S., Tarekegne, A., Babu, R., Meisel, B., et al. (2015). GENETICS.110.118521
Genetic Gains in Grain Yield through Genomic Selection in Eight Bi-parental Crossa, J., Pérez, P., de los Campos, G., Mahuku, G., Dreisigacker, S., and
Maize Populations under Drought Stress. Crop Sci. 55, 154–163. doi:10.2135/ Magorokosho, C. (2011). Genomic Selection and Prediction in Plant
CROPSCI2014.07.0460 Breeding. J. Crop Improvement 25, 239–261. doi:10.1080/15427528.2011.
Bhandari, A., Bartholomé, J., Cao-Hamadoun, T.-V., Kumari, N., Frouin, J., 558767
Kumar, A., et al. (2019). Selection of Trait-specific Markers and Multi- Crossa, J., Pérez, P., Hickey, J., Burgueño, J., Ornella, L., Cerón-Rojas, J., et al.
Environment Models Improve Genomic Predictive Ability in rice. PLOS (2014). Genomic Prediction in CIMMYTmaize andWheat Breeding Programs.
ONE 14, e0208871. doi:10.1371/JOURNAL.PONE.0208871 Heredity 112, 48–60. doi:10.1038/HDY.2013.16
Frontiers in Genetics | www.frontiersin.org 13 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
Crossa, J., Pérez-Rodríguez, P., Cuevas, J., Montesinos-López, O., Jarquín, D., de los Resistant Recombinants in Basmati rice. Plant Breed. 127, 131–139. doi:10.
Campos, G., et al. (2017). Genomic Selection in Plant Breeding: Methods, 1111/J.1439-0523.2007.01458.X
Models, and Perspectives. Trends Plant Science 22, 961–975. doi:10.1016/J. Grenier, C., Cao, T.-V., Ospina, Y., Quintero, C., Châtel, M. H., Tohme, J., et al.
TPLANTS.2017.08.011 (2015). Accuracy of Genomic Selection in a Rice Synthetic Population
Cui, Z., Dong, H., Zhang, A., Ruan, Y., He, Y., and Zhang, Z. (2020). Assessment of Developed for Recurrent Selection Breeding. PloS one 10, e0136594. doi:10.
the Potential for Genomic Selection to Improve Husk Traits in Maize. G3: 1371/JOURNAL.PONE.0136594
Genes, Genomes, Genet. 10, 3741–3749. doi:10.1534/G3.120.401600 Habier, D., Fernando, R. L., Kizilkaya, K., and Garrick, D. J. (2011). Extension of
Daetwyler, H. D., Bansal, U. K., Bariana, H. S., Hayden, M. J., and Hayes, B. J. the Bayesian Alphabet for Genomic Selection. BMC Bioinformatics 12,
(2014). Genomic Prediction for Rust Resistance in Diverse Wheat Landraces. 186–197. doi:10.1186/1471-2105-12-186/FIGURES/2
Theor. Appl. Genet. 127, 1795–1803. doi:10.1007/s00122-014-2341-8 Habier, D., Fernando, R. L., and Dekkers, J. C. M. (2009). Genomic Selection Using
Daetwyler, H. D., Hickey, J. M., Henshall, J. M., Dominik, S., Gredler, B., van der Low-Density Marker Panels. Genetics 182, 343–353. doi:10.1534/GENETICS.
Werf, J. H. J., et al. (2010). Accuracy of Estimated Genomic Breeding Values for 108.100289
Wool and Meat Traits in a Multi-Breed Sheep Population. Anim. Prod. Sci. 50, Habier, D., Fernando, R. L., and Garrick, D. J. (2013). Genomic BLUP Decoded: a
1004–1010. doi:10.1071/AN10096 Look into the Black Box of Genomic Prediction. Genetics 194, 597–607. doi:10.
Das, R. R., Vinayan, M. T., Patel, M. B., Phagna, R. K., Singh, S. B., Shahi, J. P., et al. 1534/GENETICS.113.152207
(2020). Genetic Gains with Rapid-cycle Genomic Selection for Combined Haile, T. A., Heidecker, T., Wright, D., Neupane, S., Ramsay, L., Vandenberg, A.,
Drought and Waterlogging Tolerance in Tropical maize ( Zea May S L.). et al. (2020). Genomic Selection for Lentil Breeding: Empirical Evidence. Plant
Plant Genome 13, 1–15. doi:10.1002/tpg2.20035 Genome 13, 1–30. doi:10.1002/tpg2.20002
de los Campos, G., Hickey, J. M., Pong-Wong, R., Daetwyler, H. D., and Calus,M. P. L. Hayes, B. J., Bowman, P. J., Chamberlain, A. J., and Goddard, M. E. (2009). Invited
(2013). Whole-Genome Regression and Prediction Methods Applied to Plant and Review: Genomic Selection in Dairy Cattle: Progress and Challenges. J. Dairy
Animal Breeding. Genetics 193, 327–345. doi:10.1534/GENETICS.112.143313 Sci. 92, 433–443. doi:10.3168/JDS.2008-1646
De Roos, A. P. W., Hayes, B. J., and Goddard, M. E. (2009). Reliability of Genomic Heffner, E. L., Jannink, J.-L., Sorrells, M. E., Heffner, E. L., Sorrells, M. E., Univ, C.,
Predictions across Multiple Populations. Genetics 183, 1545–1553. doi:10.1534/ et al. (2011). Genomic Selection Accuracy UsingMultifamily PredictionModels
GENETICS.109.104935 in a Wheat Breeding Program. The Plant Genome 4, 65–75. doi:10.3835/
Diaz, S., Ariza-Suarez, D., Ramdeen, R., Aparicio, J., Arunachalam, N., Hernandez, plantgenome.2010.12.0029
C., et al. (2021). Genetic Architecture and Genomic Prediction of Cooking Time Heffner, E. L., Lorenz, A. J., Jannink, J. L., and Sorrells, M. E. (2010). Plant Breeding
in Common Bean (Phaseolus vulgaris L.). Front. Plant Sci. 11, 2257. doi:10. with Genomic Selection: Gain Per Unit Time and Cost. Crop Sci. 50, 1681–1690.
3389/FPLS.2020.622213/BIBTEX doi:10.2135/cropsci2009.11.0662
dos Santos, J. P. R., Pires, L. P. M., de Castro Vasconcellos, R. C., Pereira, G. S., von Heffner, E. L., Sorrells, M. E., and Jannink, J.-L. (2009). Genomic Selection for Crop
Pinho, R. G., and Balestre, M. (2016). Genomic Selection to Resistance to Improvement. Crop Sci. 49, 1–12. doi:10.2135/CROPSCI2008.08.0512
Stenocarpella Maydis in maize Lines Using DArTseq Markers. BMC Genet. 17, Henderson, C. R., Kempthorne, O., Searle, S. R., and von Krosigk, C. M. (1959).
86. doi:10.1186/S12863-016-0392-3 The Estimation of Environmental and Genetic Trends from Records Subject to
Duangjit, J., Causse, M., and Sauvage, C. (2016). Efficiency of Genomic Selection Culling. Biometrics 15, 192. doi:10.2307/2527669
for Tomato Fruit Quality. Mol. Breed. 36, 29. doi:10.1007/S11032-016-0453-3 Holliday, J. A., Wang, T., and Aitken, S. (2012). Predicting Adaptive
Endelman, J. B. (2011). Ridge Regression and Other Kernels for Genomic Selection Phenotypes from Multilocus Genotypes in Sitka Spruce (Picea Sitchensis)
with R Package rrBLUP. The Plant Genome 4, 250–255. doi:10.3835/ Using Random Forest. Using Random For. 2, 1085–1093. doi:10.1534/g3.
PLANTGENOME2011.08.0024 112.002733
Fernando, R., and Garrick, D. (2009). GenSel- User Manual for a Portfolio of Hong, J. P., Ro, N., Lee, H. Y., Kim, G. W., Kwon, J. K., Yamamoto, E., et al.
Genomic Selection Related Analyses. Available at: http://taurus.ansci.iastate. (2020). Genomic Selection for Prediction of Fruit-Related Traits in Pepper
edu/Site/Welcome_files/GenSel%20 Manual%20v2.pdf. (Capsicum spp.). Front. Plant Sci. 11, 570871. doi:10.3389/FPLS.2020.
Fernando, R., and Grossman, M. (1989). Marker Assisted Selection Using Best 570871/BIBTEX
Linear Unbiased Prediction. Genet. Selection Evol. 21 (421), 467–477. doi:10. Howard, R., Carriquiry, A. L., and Beavis, W. D. (2014). Parametric and
1186/1297-9686-21-4-467 Nonparametric Statistical Methods for Genomic Selection of Traits with
Fikere, M., Barbulescu, D. M., Malmberg, M. M., Maharjan, P., Salisbury, P. A., Additive and Epistatic Genetic Architectures. G3: Genes, Genomes, Genet. 4,
Kant, S., et al. (2020). Genomic Prediction and Genetic Correlation of 1027–1046. doi:10.1534/G3.114.010298/-/DC1
Agronomic, Blackleg Disease, and Seed Quality Traits in Canola (Brassica Hu, Z., Li, Y., Song, X., Han, Y., Cai, X., Xu, S., et al. (2011). Genomic Value
Napus L.). Plants 9, 719–19. doi:10.3390/PLANTS9060719 Prediction for Quantitative Traits under the Epistatic Model. BMC Genet. 12,
Fones, H. N., Bebber, D. P., Chaloner, T. M., Kay, W. T., Steinberg, G., and Gurr, S. 15. doi:10.1186/1471-2156-12-15
J. (2020). Threats to Global Food Security from Emerging Fungal and Oomycete Huang, M., Balimponya, E. G., Mgonja, E. M., McHale, L. K., Luzi-Kihupi, A.,
Crop Pathogens. Nat. Food 1, 332–342. doi:10.1038/s43016-020-0075-0 Wang, G.-L., et al. (2019). Use of Genomic Selection in Breeding rice (Oryza
Friedman, J., Hastie, T., and Tibshirani, R. (2010). Regularization Paths for Sativa L.) for Resistance to rice Blast (Magnaporthe Oryzae). Mol. Breed. 39,
Generalized Linear Models via Coordinate Descent. J. Stat. Softw. 33, 1–22. 1–16. doi:10.1007/S11032-019-1023-2
doi:10.18637/jss.v033.i01 Imai, A., Kuniga, T., Yoshioka, T., Nonaka, K., Mitani, N., Fukamachi, H., et al.
Fristche-Neto, R., Akdemir, D., and Jannink, J.-L. (2018). Accuracy of Genomic (2019). Single-step Genomic Prediction of Fruit-Quality Traits Using
Selection to Predict maize Single-Crosses Obtained through Different Mating Phenotypic Records of Non-genotyped Relatives in Citrus. PLOS ONE 14,
Designs. Theor. Appl. Genet. 131, 1153–1162. doi:10.1007/s00122-018-3068-8 e0221880. doi:10.1371/JOURNAL.PONE.0221880
Furbank, R. T., and Tester, M. (2011). Phenomics - Technologies to Relieve the Jan, H. U., Abbadi, A., Lücke, S., Nichols, R. A., and Snowdon, R. J. (2016).
Phenotyping Bottleneck. Trends Plant Sci. 16, 635–644. doi:10.1016/J. Genomic Prediction of Testcross Performance in Canola (Brassica Napus).
TPLANTS.2011.09.005 PLOS ONE 11, e0147769. doi:10.1371/JOURNAL.PONE.0147769
Gianola, D., Fernando, R. L., and Stella, A. (2006). Genomic-Assisted Prediction of Janila, P., Variath, M. T., Pandey, M. K., Desmae, H., Motagi, B. N., Okori, P., et al.
Genetic Value with Semiparametric Procedures. Genetics 173, 1761–1776. (2016). Genomic Tools in Groundnut Breeding Program: Status and
doi:10.1534/GENETICS.105.049510 Perspectives. Front. Plant Sci. 7, 289. doi:10.3389/FPLS.2016.00289/BIBTEX
Gianola, D., Okut, H., Weigel, K. A., and Rosa, G. J. (2011). Predicting Complex Jeong, S., Kim, J.-Y., and Kim, N. (2020). GMStool: GWAS-Based Marker Selection
Quantitative Traits with Bayesian Neural Networks: a Case Study with Jersey Tool for Genomic Prediction from Genomic Data. Sci. Rep. 10, 1–12. doi:10.
Cows and Wheat. BMC Genet. 12, 87. doi:10.1186/1471-2156-12-87 1038/s41598-020-76759-y
Gopalakrishnan, S., Sharma, R. K., Anand Rajkumar, K., Joseph, M., Singh, V. P., Jia, Y., and Jannink, J.-L. (2012). Multiple-Trait Genomic Selection Methods
Singh, A. K., et al. (2008). Integrating Marker Assisted Background Analysis Increase Genetic Value Prediction Accuracy. Genetics 192, 1513–1522.
with Foreground Selection for Identification of superior Bacterial Blight doi:10.1534/GENETICS.112.144246
Frontiers in Genetics | www.frontiersin.org 14 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
Joukhadar, R., Thistlethwaite, R., Trethowan, R. M., Hayden, M. J., Stangoulis, J., Lozada, D. N., Mason, R. E., Sarinelli, J. M., and Brown-Guedira, G. (2019).
Cu, S., et al. (2021). Genomic Selection Can Accelerate the Biofortification of Accuracy of Genomic Selection for Grain Yield and Agronomic Traits in Soft
spring Wheat. Theor. Appl. Genet. 134, 3339–3350. doi:10.1007/S00122-021- Red winter Wheat. BMC Genet. 20, 1–12. doi:10.1186/s12863-019-0785-1
03900-4 Luan, T., Woolliams, J. A., Lien, S., Kent, M., Svendsen, M., and Meuwissen, T. H.
Juarez, M., Legua, P., Mengual, C. M., Kassem, M. A., Sempere, R. N., Gómez, P., E. (2009). The Accuracy of Genomic Selection in Norwegian Red Cattle
et al. (2013). Relative Incidence, Spatial Distribution and Genetic Diversity of Assessed by Cross-Validation. Genetics 183, 1119–1126. doi:10.1534/
Cucurbit Viruses in Eastern Spain. Ann. Appl. Biol. 162, 362–370. doi:10.1111/ GENETICS.109.107391
AAB.12029 Lush, J. L. (1943). Animal Breeding Plans. Edn 2. Charleston, South Carolina:
Juliana, P., Poland, J., Huerta-Espino, J., Shrestha, S., Crossa, J., Crespo-Herrera, L., Bibliolife DBA of Bibilio Bazaar.
et al. (2019). Improving Grain Yield, Stress Resilience and Quality of Bread Mackay, T. F. C. (2001). The Genetic Architecture of Quantitative Traits. Annu.
Wheat Using Large-Scale Genomics. Nat. Genet. 51, 1530–1539. doi:10.1038/ Rev. Genet. 35, 303–339. doi:10.1146/annurev.genet.35.102401.090633
s41588-019-0496-6 Madhavi, K. R., Rambabu, R., Abhilash Kumar, V., Vijay Kumar, S., Aruna, J.,
Juliana, P., Singh, R. P., Singh, P. K., Crossa, J., Rutkoski, J. E., Poland, J. A., et al. Ramesh, S., et al. (2016). Marker Assisted Introgression of Blast (Pi-2 and Pi-54)
(2017). Comparison of Models and Whole-Genome Profiling Approaches for Genes in to the Genetic Background of Elite, Bacterial Blight Resistant Indica
Genomic-Enabled Prediction of Septoria Tritici Blotch, Stagonospora rice Variety, Improved Samba Mahsuri. Euphytica 212, 331–342. doi:10.1007/
Nodorum Blotch, and Tan Spot Resistance in Wheat. Plant Genome 10, S10681-016-1784-1
1–16. doi:10.3835/PLANTGENOME2016.08.0082 Maenhout, S., De Baets, B., Haesaert, G., and Van Bockstaele, E. (2007). Support
Klápště, J., Dungey, H. S., Telfer, E. J., Suontama, M., Graham, N. J., Li, Y., et al. Vector Machine Regression for the Prediction of maize Hybrid Performance.
(2020). Marker Selection in Multivariate Genomic Prediction Improves Theor. Appl. Genet. 115, 1003–1013. doi:10.1007/s00122-007-0627-9
Accuracy of Low Heritability Traits. Front. Genet. 11, 499094. doi:10.3389/ Majumdar, S. G., Rai, A., and Mishra, D. C. (2020). Integrated Framework for
FGENE.2020.499094/FULL Selection of Additive and Nonadditive Genetic Markers for Genomic Selection.
Krishnan, S. G., Singh, A. K., Rathour, R., Nagarajan, M., Bhowmick, P. K., Ellur, R. J. Comput. Biol. 27, 845–855. doi:10.1089/CMB.2019.0223
K., et al. (2019). Rice Variety Pusa Samba 1850. Indian J. Genet. 79, 109–110. Majumdar, S. G., Rai, A., and Mishra, D. C. (2019). Package ‘GSelection’, 1–14.
Kumar, S., Chagné, D., Bink, M. C. A. M., Volz, R. K., Whitworth, C., and Carlisle, Available at: https://rdrr.io/cran/GSelection/man/GSelection-package.html.
C. (2012). Genomic Selection for Fruit Quality Traits in Apple Mangin, B., Bonnafous, F., Blanchet, N., Boniface, M.-C., Bret-Mestries, E.,
(Malus×domestica Borkh.). PLoS One 7, e36674. doi:10.1371/JOURNAL. Carrère, S., et al. (2017). Genomic Prediction of sunflower Hybrids Oil
PONE.0036674 Content. Front. Plant Sci. 8, 1633. doi:10.3389/FPLS.2017.01633/
Kumar, S., Kirk, C., Deng, C. H., Shirtliff, A., Wiedow, C., Qin, M., et al. (2019). BIBTEX
Marker-trait Associations and Genomic Predictions of Interspecific Pear Marulanda, J. J., Mi, X., Melchinger, A. E., Xu, J.-L., Würschum, T., and Longin, C.
(Pyrus) Fruit Characteristics. Sci. Rep. 9, 1–10. doi:10.1038/s41598-019- F. H. (2016). Optimum Breeding Strategies Using Genomic Selection for
45618-w Hybrid Breeding in Wheat, maize, rye, Barley, rice and Triticale. Theor.
Lam, H.-M., Coschigano, K. T., Oliveira, I. C., Melo-Oliveira, R., and Coruzzi, G. Appl. Genet. 129, 1901–1913. doi:10.1007/s00122-016-2748-5
M. (1996). The Molecular-Genetics of Nitrogen Assimilation into Amino Acids Meuwissen, T. H. (2009). Accuracy of Breeding Values of ’unrelated’ Individuals
in Higher Plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 47, 569–593. doi:10. Predicted by Dense SNP Genotyping. Genet. Sel Evol. 41, 35–39. doi:10.1186/
1146/annurev.arplant.47.1.569 1297-9686-41-35/TABLES/3
Legarra, A., and Reverter, A. (2018). Semi-parametric Estimates of Population Meuwissen, T. H. E., Hayes, B. J., and Goddard, M. E. (2001). Prediction of Total
Accuracy and Bias of Predictions of Breeding Values and Future Phenotypes Genetic Value Using Genome-wide Dense Marker Maps. Genetics 157,
Using the LR Method. Genet. Sel Evol. 50, 53–18. doi:10.1186/S12711-018- 1819–1829. doi:10.1093/GENETICS/157.4.1819
0426-6/FIGURES/3 Michel, S., Löschenberger, F., Ametz, C., Pachler, B., Sparry, E., and Bürstmayr, H.
Li, Y., Ruperao, P., Batley, J., Edwards, D., Khan, T., Colmer, T. D., et al. (2018). (2019). Combining Grain Yield, Protein Content and Protein Quality by Multi-
Investigating Drought Tolerance in Chickpea Using Genome-wide Association Trait Genomic Selection in Bread Wheat. Theor. Appl. Genet. 132, 2767–2780.
Mapping and Genomic Selection Based onWhole-Genome Resequencing Data. doi:10.1007/S00122-019-03386-1
Front. Plant Sci. 9, 190. doi:10.3389/FPLS.2018.00190/BIBTEX Minamikawa, M. F., Nonaka, K., Kaminuma, E., Kajiya-Kanegae, H., Onogi, A.,
Liu, B., Asseng, S., Müller, C., Ewert, F., Elliott, J., Lobell, D. B., et al. (2016). Similar Goto, S., et al. (2017). Genome-wide Association Study and Genomic Prediction
Estimates of Temperature Impacts on Global Wheat Yield by Three in Citrus: Potential of Genomics-Assisted Breeding for Fruit Quality Traits. Sci.
Independent Methods. Nat. Clim Change 6, 1130–1136. doi:10.1038/ Rep. 7, 1–13. doi:10.1038/s41598-017-05100-x
NCLIMATE3115 Mirdita, V., He, S., Zhao, Y., Korzun, V., Bothe, R., Ebmeyer, E., et al. (2015).
Liu, X., Wang, H., Wang, H., Guo, Z., Xu, X., Liu, J., et al. (2018). Factors Affecting Potential and Limits of Whole Genome Prediction of Resistance to Fusarium
Genomic Selection Revealed by Empirical Evidence in maize. Crop J. 6, Head Blight and Septoria Tritici Blotch in a Vast Central European Elite winter
341–352. doi:10.1016/J.CJ.2018.03.005 Wheat Population. Theor. Appl. Genet. 128, 2471–2481. doi:10.1007/S00122-
Long, N., Gianola, D., Rosa, G. J. M., and Weigel, K. A. (2011). Application of 015-2602-1
Support Vector Regression to Genome-Assisted Prediction of Quantitative Mishra, D. C., Budhlakoti, N., Majumdar, S. G., and Rai, A. (2021). Innovations
Traits. Theor. Appl. Genet. 123, 1065–1074. doi:10.1007/S00122-011-1648-Y in Genomic Selection : Statistical Perspective. 101–111. Available at: https://
Longin, C. F. H., Reif, J. C., and Würschum, T. (2014). Long-term Perspective of ssca.org.in/media/9_Spl_Proceedings_2021_006072021_Dwijesh_Mishra_
Hybrid versus Line Breeding in Wheat Based on Quantitative Genetic Theory. Final.pdf.
Theor. Appl. Genet. 127, 1635–1641. doi:10.1007/S00122-014-2325-8 Mohan, M., Nair, S., Bhagwat, A., Krishna, T. G., Yano, M., Bhatia, C. R., et al.
Lorenz, A. J., Chao, S., Asoro, F. G., Heffner, E. L., Hayashi, T., Iwata, H., et al. (1997). Genome Mapping, Molecular Markers and Marker-Assisted Selection
(2011). Genomic Selection in Plant Breeding. Adv. Agron. 110, 77–123. doi:10. in Crop Plants. Mol. Breed. 3, 87–103. doi:10.1023/A:1009651919792
1016/B978-0-12-385531-2.00002-5 Moore, J. H., andWilliams, S. M. (2009). Epistasis and its Implications for Personal
Lorenz, A. J., Smith, K. P., and Jannink, J. L. (2012). Potential and Optimization of Genetics. Am. J. Hum. Genet. 85, 309–320. doi:10.1016/J.AJHG.2009.08.006
Genomic Selection for Fusarium Head Blight Resistance in Six-Row Barley. Nakaya, A., and Isobe, S. N. (2012). Will Genomic Selection Be a Practical Method
Crop Sci. 52, 1609–1621. doi:10.2135/cropsci2011.09.0503 for Plant Breeding? Ann. Bot. 110, 1303–1316. doi:10.1093/AOB/MCS109
Lorenzana, R. E., and Bernardo, R. (2009). Accuracy of Genotypic Value Neeraja, C. N., Maghirang-Rodriguez, R., Pamplona, A., Heuer, S., Collard, B. C. Y.,
Predictions for Marker-Based Selection in Biparental Plant Populations. Septiningsih, E. M., et al. (2007). A Marker-Assisted Backcross Approach for
Theor. Appl. Genet. 120, 151–161. doi:10.1007/s00122-009-1166-3 Developing Submergence-Tolerant rice Cultivars. Theor. Appl. Genet. 115 (6),
Lozada, D. N., and Carter, A. H. (2020). Genomic Selection in Winter Wheat 767–776. doi:10.1007/s00122-007-0607-0
Breeding Using a Recommender Approach. Genes 11, 1–14. doi:10.3390/ Norman, A., Taylor, J., Edwards, J., and Kuchel, H. (2018). Optimising Genomic
GENES11070779 Selection in Wheat: Effect of Marker Density, Population Size and Population
Frontiers in Genetics | www.frontiersin.org 15 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
Structure on Prediction Accuracy. G3 Genes|Genomes|Genetics 8, 2889–2899. Sallam, A. H., and Smith, K. P. (2016). Genomic Selection Performs Similarly to
doi:10.1534/G3.118.200311 Phenotypic Selection in Barley. Crop Sci. 56, 2871–2881. doi:10.2135/
Nsibi, M., Gouble, B., Bureau, S., Flutre, T., Sauvage, C., Audergon, J.-M., et al. CROPSCI2015.09.0557
(2020). Adoption and Optimization of Genomic Selection to Sustain Breeding Sarinelli, J. M., Murphy, J. P., Tyagi, P., Holland, J. B., Johnson, J.W., Mergoum,M.,
for Apricot Fruit Quality. G3 Genes|Genomes|Genetics 10, 4513–4529. doi:10. et al. (2019). Training Population Selection and Use of Fixed Effects to Optimize
1534/G3.120.401452 Genomic Predictions in a Historical USA winter Wheat Panel. Theor. Appl.
Ogutu, J. O., Schulz-Streeck, T., and Piepho, H.-P. (2012). Genomic Selection Genet. 132, 1247–1261. doi:10.1007/S00122-019-03276-6
Using Regularized Linear Regression Models: ridge Regression, Lasso, Elastic Schmidt, M., Kollers, S., Maasberg-Prelle, A., Großer, J., Schinkel, B., Tomerius, A.,
Net and Their Extensions. BMC Proc. 6, S10. doi:10.1186/1753-6561-6-S2-S10 et al. (2016). Prediction of Malting Quality Traits in Barley Based on Genome-
Onogi, A., Ideta, O., Inoshita, Y., Ebana, K., Yoshioka, T., Yamasaki, M., et al. wide Marker Data to Assess the Potential of Genomic Selection. Theor. Appl.
(2015). Exploring the Areas of Applicability of Whole-Genome Prediction Genet. 129, 203–213. doi:10.1007/s00122-015-2639-1
Methods for Asian rice (Oryza Sativa L.). Theor. Appl. Genet. 128, 41–53. doi:10. Sehgal, D., Rosyara, U., Mondal, S., Singh, R., Poland, J., and Dreisigacker, S.
1007/S00122-014-2411-Y (2020). Incorporating Genome-wide Association Mapping Results into
Pandey, M. K., Chaudhari, S., Jarquin, D., Janila, P., Crossa, J., Patil, S. C., et al. Genomic Prediction Models for Grain Yield and Yield Stability in CIMMYT
(2020). Genome-based Trait Prediction in Multi- Environment Breeding Trials Spring Bread Wheat. Front. Plant Sci. 11, 197. doi:10.3389/FPLS.2020.00197
in Groundnut. Theor. Appl. Genet. 133, 3101–3117. doi:10.1007/S00122-020- Semagn, K., Beyene, Y., Warburton, M. L., Tarekegne, A., Mugo, S., Meisel, B., et al.
03658-1/TABLES/5 (2013). Meta-analyses of QTL for Grain Yield and Anthesis Silking Interval in
Pérez, P., and de los Campos, G. (2014). Genome-Wide Regression and Prediction 18 maize Populations Evaluated under Water-Stressed and Well-Watered
with the BGLR Statistical Package. Genetics 198, 483–495. doi:10.1534/genetics. Environments. BMC Genomics 14, 313–316. doi:10.1186/1471-2164-14-313/
114.164442 TABLES/4
Pérez-Rodríguez, P., Gianola, D., González-Camacho, J. M., Crossa, J., Manès, Y., Shikha, M., Kanika, A., Rao, A. R., Mallikarjuna, M. G., Gupta, H. S., and Nepolean,
and Dreisigacker, S. (2012). Comparison between Linear and Non-parametric T. (2017). Genomic Selection for Drought Tolerance Using Genome-wide SNPs
Regression Models for Genome-Enabled Prediction in Wheat. G3: Genes, in Maize. Front. Plant Sci. 8, 1–12. doi:10.3389/fpls.2017.00550
Genomes, Genet. 2, 1595–1605. doi:10.1534/G3.112.003665/-/DC1 Singh, A. K., Gopala Krishnan, S., Ellur, R. K., Bhowmick, P. K., Nagarajan, M.,
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., Heisterkamp, S., and VanWilligen, B. Vinod, K. K., et al. (2017a). Notification of Basmati rice Variety, Pusa Basmati
(2017). Package ‘nlme’. Available at: https://cran.r-project.org/web/packages/ 1728. Indian J. Genet. 77, 584.
nlme/nlme.pdf. Singh, A. K., Gopala Krishnan, S., Nagarajan, M., Bhowmick, P. K., Ellur, R. K.,
Poland, J., Endelman, J., Dawson, J., Rutkoski, J., Wu, S., Manes, Y., et al. (2012). Haritha, B., et al. (2017b). Notification of Basmati rice Variety Pusa Basmati
Genomic Selection inWheat Breeding Using Genotyping-by-Sequencing. Plant 1637. Indian J. Genet. 77, 583–584.
Genome 5, 1–11. doi:10.3835/PLANTGENOME2012.06.0006 Spindel, J., Begum, H., Akdemir, D., Virk, P., Collard, B., Redoña, E., et al. (2015).
Qin, F., Shinozaki, K., and Yamaguchi-Shinozaki, K. (2011). Achievements and Correction: Genomic Selection and Association Mapping in Rice (Oryza
Challenges in Understanding Plant Abiotic Stress Responses and Tolerance. Sativa): Effect of Trait Genetic Architecture, Training Population
Plant Cel Physiol. 52, 1569–1582. doi:10.1093/PCP/PCR106 Composition, Marker Number and Statistical Model on Accuracy of Rice
Rai, K. N., Hash, C. T., Singh, A. K., and Velu, G. (2008). Adaptation and Quality Genomic Selection in Elite, Tropical Rice Breeding Lines. Plos Genet. 11,
Traits of a Germplasm-Derived Commercial Seed Parent of Pearl Millet. Plant e1005350. doi:10.1371/JOURNAL.PGEN.1005350
Genet. Resour. Newsl. 154, 20–24. Stewart-Brown, B. B., Song, Q., Vaughn, J. N., and Li, Z. (2019). Genomic Selection
Reif, J. C., Zhao, Y., Würschum, T., Gowda, M., and Hahn, V. (2013). Genomic for Yield and Seed Composition Traits within an Applied Soybean Breeding
Prediction of sunflower Hybrid Performance. Plant Breed 132, 107–114. doi:10. Program. G3 Genes, Genomes, Genet. 9, 2253–2265. doi:10.1534/g3.118.200917
1111/pbr.12007 Sukumaran, S., Jarquin, D., Crossa, J., and Reynolds, M. (2018). Genomic-enabled
Reynolds, M. P., and Ortiz, R. (2010). “Adapting Crops to Climate Change: a Prediction Accuracies Increased by Modeling Genotype × Environment
Summary,” in Climate Change and Crop Production. Editor Interaction in Durum Wheat. Plant Genome 11, 170112. doi:10.3835/
M. P. Reynolds (Wallingford,UK: CABI), 1–8. doi:10.1079/ PLANTGENOME2017.12.0112
9781845936334.0001 Sun, J., Poland, J. A., Mondal, S., Crossa, J., Juliana, P., Singh, R. P., et al. (2019).
Ribaut, J.-M., and Ragot, M. (2007). Marker-assisted Selection to Improve Drought High-throughput Phenotyping Platforms Enhance Genomic Selection for
Adaptation in maize: the Backcross Approach, Perspectives, Limitations, and Wheat Grain Yield across Populations and Cycles in Early Stage. Theor.
Alternatives. J. Exp. Bot. 58, 351–360. doi:10.1093/JXB/ERL214 Appl. Genet. 132, 1705–1720. doi:10.1007/S00122-019-03309-0
Rio, S., Mary-Huard, T., Moreau, L., and Charcosset, A. (2019). Genomic Selection Sun, J., Rutkoski, J. E., Poland, J. A., Crossa, J., Jannink, J. L., and Sorrells, M. E.
Efficiency and A Priori Estimation of Accuracy in a Structured Dent maize (2017). Multitrait, Random Regression, or Simple Repeatability Model in High-
Panel. Theor. Appl. Genet. 132, 81–96. doi:10.1007/s00122-018-3196-1 Throughput Phenotyping Data Improve Genomic Prediction for Wheat Grain
Roth, M., Muranty, H., Di Guardo, M., Guerra, W., Patocchi, A., and Costa, F. Yield. Plant Genome 10, 1–12. doi:10.3835/PLANTGENOME2016.11.0111
(2020). Genomic Prediction of Fruit Texture and Training Population Tanaka, E. (2020). Simple Outlier Detection for a Multi-environmental Field Trial.
Optimization towards the Application of Genomic Selection in Apple. Biometrics 76, 1374–1382. doi:10.1111/BIOM.13216
Hortic. Res. 7, 1–14. doi:10.1038/s41438-020-00370-5 Tanaka, E. (2018). Simple Robust Genomic Prediction and Outlier Detection for a
Rothman, A. J., Levina, E., and Zhu, J. (2010). Sparse Multivariate Regression with Multi-Environmental Field Trial. arXiv preprint arXiv:1807.07268, 1–25.
Covariance Estimation. J. Comput. Graphical Stat. 19, 947–962. doi:10.1198/ Tester, M., and Langridge, P. (2010). Breeding Technologies to Increase Crop
JCGS.2010.09188 Production in a Changing World. Science 327, 818–822. doi:10.1126/SCIENCE.
Rutkoski, J., Benson, J., Jia, Y., Brown-Guedira, G., Jannink, J.-L., and Sorrells, M. 1183700
(2012). Evaluation of Genomic Prediction Methods for Fusarium Head Blight Tibshirani, R. (1996). Regression Shrinkage and Selection via the Lasso. J. R. Stat.
Resistance in Wheat. The Plant Genome 5, 51–61. doi:10.3835/ Soc. Ser. B (Methodological) 58, 267–288. doi:10.1111/J.2517-6161.1996.
PLANTGENOME2012.02.0001 TB02080.X
Rutkoski, J., Poland, J., Mondal, S., Autrique, E., Pérez, L. G., Crossa, J., et al. (2016). Tiede, T., and Smith, K. P. (2018). Evaluation and Retrospective Optimization of
Canopy Temperature and Vegetation Indices from High-Throughput Genomic Selection for Yield and Disease Resistance in spring Barley. Mol.
Phenotyping Improve Accuracy of Pedigree and Genomic Selection for Breed. 38, 55. doi:10.1007/S11032-018-0820-3
Grain Yield in Wheat. G3: Genes, Genomes, Genet. 6, 2799–2808. doi:10. Usai, M. G., Goddard, M. E., and Hayes, B. J. (2009). LASSO with Cross-Validation
1534/G3.116.032888 for Genomic Selection. Genet. Res. 91, 427–436. doi:10.1017/
Sallam, A. H., Endelman, J. B., Jannink, J. L., and Smith, K. P. (2015). Assessing S0016672309990334
Genomic Selection Prediction Accuracy in a Dynamic Barley Breeding VanRaden, P. M. (2008). Efficient Methods to Compute Genomic Predictions.
Population. Plant Genome 8, 20. doi:10.3835/PLANTGENOME2014.05.0020 J. Dairy Sci. 91, 4414–4423. doi:10.3168/JDS.2007-0980
Frontiers in Genetics | www.frontiersin.org 16 February 2022 | Volume 13 | Article 832153
Budhlakoti et al. A Comprehensive Review on Genomic Selection
Varshney, R. K., Mohan, S. M., Gaur, P. M., Chamarthi, S. K., Singh, V. K., Yabe, S., Yoshida, H., Kajiya-Kanegae, H., Yamasaki, M., Iwata, H., Ebana, K., et al.
Srinivasan, S., et al. (2014a). Marker-Assisted Backcrossing to Introgress (2018). Description of Grain Weight Distribution Leading to Genomic
Resistance to Fusarium Wilt Race 1 and Ascochyta Blight in C 214, an Elite Selection for Grain-Filling Characteristics in rice. PLOS ONE 13, e0207627.
Cultivar of Chickpea. Plant Genome 7, 35. doi:10.3835/plantgenome2013.10. doi:10.1371/JOURNAL.PONE.0207627
0035 Yuan, Y., Cairns, J. E., Babu, R., Gowda, M., Makumbi, D., Magorokosho, C., et al.
Varshney, R. K., Pandey, M. K., Janila, P., Nigam, S. N., Sudini, H., Gowda, M. V. C., (2019). Genome-wide Association Mapping and Genomic Prediction Analyses
et al. (2014b). Marker-assisted Introgression of a QTL Region to Improve Rust Reveal the Genetic Architecture of Grain Yield and Flowering Time under
Resistance in Three Elite and Popular Varieties of Peanut (Arachis hypogaea L.). Drought and Heat Stress Conditions in maize. Front. Plant Sci. 9, 1919. doi:10.
Theor. Appl. Genet. 127, 1771–1781. doi:10.1007/S00122-014-2338-3 3389/FPLS.2018.01919/FULL
Varshney, R. K., Singh, V. K., Kumar, A., Powell, W., and Sorrells, M. E. (2018). Zhang, H., Yin, L., Wang, M., Yuan, X., and Liu, X. (2019). Factors Affecting the
Can Genomics Deliver Climate-Change Ready Crops? Curr. Opin. Plant Biol. Accuracy of Genomic Selection for Agricultural Economic Traits in maize,
45, 205–211. doi:10.1016/J.PBI.2018.03.007 Cattle, and Pig Populations. Front. Genet. 10, 189. doi:10.3389/FGENE.2019.
Váry, Z., Mullins, E., Mcelwain, J. C., and Doohan, F. M. (2015). The Severity of Wheat 00189/BIBTEX
Diseases Increases when Plants and Pathogens Are Acclimatized to Elevated Carbon Zhang, X., Pérez-Rodríguez, P., Burgueño, J., Olsen, M., Buckler, E., Atlin, G., et al.
Dioxide. Glob. Change Biol. 21, 2661–2669. doi:10.1111/GCB.12899 (2017). Rapid Cycling Genomic Selection in a Multiparental Tropical maize
Vasistha, N. K., Balasubramaniam, A., Mishra, V. K., Srinivasa, J., Chand, R., and Population. G3: Genes, Genomes, Genet. 7, 2315–2326. doi:10.1534/G3.117.
Joshi, A. K. (2017). Molecular Introgression of Leaf Rust Resistance Gene Lr34 043141
Validates Enhanced Effect on Resistance to Spot Blotch in spring Wheat. Zhao, Y., Gowda, M., Liu, W., Würschum, T., Maurer, H. P., Longin, F. H., et al.
Euphytica 213, 1–10. doi:10.1007/s10681-017-2051-9 (2012). Accuracy of Genomic Selection in European maize Elite Breeding
Vazquez, A. I., Bates, D. M., Rosa, G. J. M., Gianola, D., and Weigel, K. A. (2010). Populations. Theor. Appl. Genet. 124, 769–776. doi:10.1007/S00122-011-
Technical Note: An R Package for Fitting Generalized Linear Mixed Models in 1745-Y
Animal Breeding1. J. Anim. Sci. 88, 497–504. doi:10.2527/JAS.2009-1952 Zhao, Y., Mette, M. F., and Reif, J. C. (2015). Genomic Selection in Hybrid
Viswanatha, K. P., Patil, R., Upadhyaya, H. D., Khan, H., Gururaj, S., ., S., et al. Breeding. Plant Breed 134, 1–10. doi:10.1111/PBR.12231
(2020). Genetic Diversity, Association and Principle Component Analyses for Zhao, Y., Zeng, J., Fernando, R., and Reif, J. C. (2013). Genomic Prediction of
Agronomical and Quality Traits in Genomic Selection Training Population of Hybrid Wheat Performance. Crop Sci. 53, 802–810. doi:10.2135/
Groundnut (Arachis hypogaea L.). Ijgpb 80, 282–290. doi:10.31742/IJGPB.80. CROPSCI2012.08.0463
3.7 Ziyatdinov, A., Ve1;zquez-Santiago, M., Brunel, H., Martinez-Perez, A., Aschard,
Vivek, B. S., Krishna, G. K., Vengadessan, V., Babu, R., Zaidi, P. H., Kha, L. Q., et al. H., and Soria, J. M. (2018). lme4qtl: Linear Mixed Models With Flexible
(2017). Use of Genomic Estimated Breeding Values Results in Rapid Genetic Covariance Structure For Genetic Studies Of Related Individuals. BMC
Gains for Drought Tolerance in Maize. Plant Genome 10. doi:10.3835/ Bioinformat. 19, 68. doi:10.1186/s12859-018-2057-x
PLANTGENOME2016.07.0070/FORMAT/PDF Zou, H., and Hastie, T. (2005). Regularization and Variable Selection via the Elastic
Wang, X., Xu, Y., Hu, Z., and Xu, C. (2018). Genomic Selection Methods for Crop Net. J. R. Stat. Soc B 67, 301–320. doi:10.1111/j.1467-9868.2005.00503.x
Improvement: Current Status and Prospects. Crop J. 6, 330–340. doi:10.1016/j.
cj.2018.03.001 Conflict of Interest: The authors declare that the research was conducted in the
Werner, C. R., Voss-Fels, K. P., Miller, C. N., Qian, W., Hua, W., Guan, C. Y., et al. absence of any commercial or financial relationships that could be construed as a
(2018). Effective Genomic Selection in a Narrow-Genepool Crop with Low- potential conflict of interest.
Density Markers: Asian Rapeseed as an Example. Plant Genome 11, 170084.
doi:10.3835/plantgenome2017.09.0084 Publisher’s Note: All claims expressed in this article are solely those of the authors
WHO/FAO (2003). Diet, Nutrition and the Prevention of Chronic Diseases: and do not necessarily represent those of their affiliated organizations, or those of
Recommendations for Preventing Excess Weight Gains and Obesity. Geneva, the publisher, the editors, and the reviewers. Any product that may be evaluated in
Switzerland: WHO, 1–148. this article, or claim that may be made by its manufacturer, is not guaranteed or
Xiong, S., Wang, M., Zou, J., Meng, J., and Liu, Y. (2020). A Two-Stage Method for endorsed by the publisher.
Improving the Prediction Accuracy of Complex Traits by Incorporating
Genotype by Environment Interactions inBrassica Napus. Discrete Dyn. Nat. Copyright © 2022 Budhlakoti, Kushwaha, Rai, Chaturvedi, Kumar, Pradhan,
Soc. 2020, 1–12. doi:10.1155/2020/7959508 Kumar, Kumar, Juliana, Mishra and Kumar. This is an open-access article
Xu, S. (2007). An Empirical Bayes Method for Estimating Epistatic Effects of distributed under the terms of the Creative Commons Attribution License (CC
Quantitative Trait Loci. Biometrics 63, 513–521. doi:10.1111/J.1541-0420.2006. BY). The use, distribution or reproduction in other forums is permitted, provided the
00711.X original author(s) and the copyright owner(s) are credited and that the original
Xu, Y., Wang, X., Ding, X., Zheng, X., Yang, Z., Xu, C., et al. (2018). Genomic publication in this journal is cited, in accordance with accepted academic practice.
Selection of Agronomic Traits in Hybrid rice Using anNCII Population. Rice (N No use, distribution or reproduction is permitted which does not comply with
Y) 11, 32–10. doi:10.1186/S12284-018-0223-4/FIGURES/5 these terms.
Frontiers in Genetics | www.frontiersin.org 17 February 2022 | Volume 13 | Article 832153