Assessment of molecular markers and marker-assisted selection for drought tolerance in barley (Hordeum vulgare L.)

Akmaral Baidyussen ,Gulmira Khassanova ,Maral Utebayev ,Satyvaldy Jatayev ,Rystay Kushanova ,Sholpan Khalbayeva,Aigul Amangeldiyeva,Raushan Yerzhebayeva,Kulpash Bulatova,Carly Schramm,Peter Anderson,Colin L.D.Jenkins,Kathleen L.Soole,Yuri Shavrukov#

1 Faculty of Agronomy,S. Seifullin Kazakh AgroTechnical Research University,Astana 010000,Kazakhstan

2 A.I. Barayev Research and Production Centre of Grain Farming,Shortandy 021601,Kazakhstan

3 Kazakh Research Institute of Agriculture and Plant Growing,Almalybak,Almaty District 040909,Kazakhstan

4 College of Science and Engineering,Biological Sciences,Flinders University,Adelaide,SA 5042,Australia

Abstract This review updates the present status of the field of molecular markers and marker-assisted selection (MAS),using the example of drought tolerance in barley.The accuracy of selected quantitative trait loci (QTLs),candidate genes and suggested markers was assessed in the barley genome cv.Morex.Six common strategies are described for molecular marker development,candidate gene identification and verification,and their possible applications in MAS to improve the grain yield and yield components in barley under drought stress.These strategies are based on the following five principles: (1) Molecular markers are designated as genomic ‘tags’,and their ‘prediction’ is strongly dependent on their distance from a candidate gene on genetic or physical maps;(2) plants react differently under favourable and stressful conditions or depending on their stage of development;(3) each candidate gene must be verified by confirming its expression in the relevant conditions,e.g.,drought;(4) the molecular marker identified must be validated for MAS for tolerance to drought stress and improved grain yield;and (5) the small number of molecular markers realized for MAS in breeding,from among the many studies targeting candidate genes,can be explained by the complex nature of drought stress,and multiple stress-responsive genes in each barley genotype that are expressed differentially depending on many other factors.

Keywords: barley,candidate genes,drought tolerance,gene verification via expression,grain yield,marker-assisted selection (MAS),molecular markers,quantitative trait loci (QTLs),strategy for MAS

1.lntroduction

Traditional plant breeding is based on observations of individual plant phenotypes during their cycles of propagation,or during the segregation of hybrid populations(Schmid and Thorwarth 2014).Not surprisingly,the current transition from conventional breeding to modern molecular technology,which uses additional knowledge about plant genomes,genes and genotyping,can help breeders to speed up selection more effectively.However,most of our modern molecular technologies are so complex that their use is restricted to academic pursuits or for the strategic development of molecular genetics in plants.In contrast,simple molecular methods are in strong demand for re-routing the selection of genotypes with desired traits,and the elimination of disadvantageous lines.The expectation is that such requests from breeders will be adequately addressed by molecular laboratories within a breeding company,a nearby university,or a collaborative research centre (Cobbetal.2019).

Molecular markers are regarded as the best option for answering these breeders’ requests.Using simple molecular techniques,these markers can be used as genetic ‘flags’which indicate preferable genotypes for selection.Many different types of molecular markers are currently available,and they have been developed,verified,and applied towards various aims,including plant genotyping (Lietal.2007;Boopathi 2020).Despite their differing natures and detection systems,all molecular markers are based on identified genomic regions in chromosomes that show significant associations with some studied traits.

The ultimate aim of molecular markers is their practical application in marker-assisted selection (MAS),where preferable genotypes can be selected and unwanted genotypes discarded based on the molecular marker scores (Mohanetal.1997;Chopra 2014;Schmid and Thorwarth 2014;Sallametal.2015;Kumaretal.2020).Therefore,molecular markers can greatly assist breeding,but their adoption is dependent on the complexity of the technology,which in many cases is beyond the capabilities of the users.

The efficiency of MAS is directly related to the type and number of molecular markers used,the genotypes analysed,and further selection criteria specific to the particular breeding program.The theoretical background of MAS was described thoroughly in an earlier report(Lande and Thompson 1990).

The application of molecular markers in MAS is a very attractive option.Ideally,the molecular markers are simple to use,and young seedlings that are early in development can be used for the molecular analyses.MAS is then efficient in space and other resources since only a limited number of selected genotypes need to be grown to harvest.The expenses for field trials in breeding can be reduced significantly,with simultaneous enrichment of genotypes with preferable traits (Chopra 2014;Schmid and Thorwarth 2014).There are many successful examples of MAS application in various crops,including barley.For example,MAS with marker954-1377located in the4HG0417270gene(MLOC_53422),which encodes a lipid-transfer protein,was successful for breeding of barley genotypes with either erect or prostrate growth habits (Mikołajczaketal.2016).

However,working with molecular markers and MAS also presents certain limitations and challenges.There are several major issues which must be considered.

(1) Molecular markers are very diverse,but they are,as the term implies,only ‘markers’.They simply flag that‘an important gene is somewhere close to the marker’.However,the accuracy of the resulting marker ‘prediction’is strongly dependent on its distance from the candidate gene on the genetic or physical map.

(2) During development,plants respond differently to favourable and stressful conditions.Therefore,the effectiveness of molecular markers and any candidate genes will depend on the developmental stages of the plants and their reaction to the specific environment.

(3) The expression of any candidate gene and its association with a particular environmental stress,such as drought,must be verified.

(4) Ideally,the identified molecular marker must be confirmed using MAS for tolerance to abiotic stress and improved grain yield (GY) in field trials,or in controlled conditions with simulated stress.

(5) Finally,there is a paradox that must be answered.With the availability of so many molecular markers and target genes in the published literature,why are there so few recommendations for MAS in plant breeding? Why have so few studies verified the functional analysisviathe expression of the targeted candidate genes?

The present review aims to address these issues using the example of barley (HordeumvulgareL.),specifically targeting the improvement of drought tolerance,as well as better GY and yield components,in dry conditions.We have conducted a detailed investigation of suggested markers to assess whether they are accurate and helpful,by aligning them on the genome sequence of barleycv.Morex,and we also determined how close they are to their practical application in MAS in barley breeding.This review is designed to be understood by anyone with an interest in this area,from amateurs to academics,researchers and breeders.Additionally,the supplementary materials were prepared to provide detailed information on the published molecular markers,their sequences,genetic mapping,the identified corresponding candidate genes and genetic regions in the barley genome and their positions on the physical map.

2.General statements

2.1.Plant responses to drought are complex and variable

Drought stress results in many adaptive physiological and biochemical changes in plants to minimize water loss,including turgor loss and adjusted osmotic status,deeper root growth,stomatal closure,reduction of leaf area and leaf senescence (Hu and Xiong 2014;Kebedeetal.2019;Alietal.2021).The main challenges in understanding plant drought tolerance are the diversity of plant reactions and the large pool of droughtresponsive genes that come into play.Causative factors relating to the drought itself include the period of onset(early,middle or terminal drought,depending on plant developmental stage),severity (mild,medium or severe),duration (short,medium or long),single or multiple‘waves’ (with the possibility for plants to intermittently recover),and potential interactions with other stresses.Factors relating to the plants include firstly,their genotypes,developmental stages (from germinating seeds through plant growth to the reproductive stage and final maturation and harvest),and their interaction with the environment.Therefore,the complexity of plant reactions to drought is the outcome and interplay of many different components (Kebedeetal.2019;Alietal.2021;Maroketal.2021).

Most researchers have imposed the direct application of drought by withdrawing the supply of water to plants in pots with soil,or by rain-out shelters in field trials.This type of drought is more ‘natural’ and is related to the reduction of water (or moisture) capacity in the soil or substrate mix.However,other indirect methods to simulate drought have sometimes been used,such as in hydroponics with growth solution;the addition of substances to absorb water like polyethylene glycol(PEG)viareduced osmotic potential;and dehydration of the root system.In addition,foliar sprays with desiccants like potassium iodide (KI) simulate a drought effectviachlorophyll loss and the start of early leaf senescence without any changes to the water capacity in the soil or osmotic potential in roots (Kalladanetal.2013;Tarawnehetal.2020).However,the results from simulated methods used as preliminary tests need to be verified in real field trials with more natural drought conditions.

2.2.Barley cv.Morex,the most convenient reference genotype with a fully sequenced genome

Cultivated barley (HordeumvulgareL.) is an important cereal crop used for the brewing,food and animal feed industries.A diploid plant species with seven pairs of chromosomes,it has a moderate genome size of about 5.3 Gb.Fully sequenced genomes of several cultivars are publicly available including Bowman and Barke(IBGSC 2012).However,in order to provide a standard single format presentation in this review,any references to marker locations and genetic loci are based only on thecv.Morex genome sequence,V.3 (Mascheretal.2021),accessedviathe Gramene website (www.gramene.org).Compared to other cereals,barley is known to be well-adapted to different levels of drought,with various morphological,biochemical and molecular mechanisms of drought tolerance (Maroketal.2021;Fatemietal.2022).However,it should be noted that the reactions of barley genotypes can be very different in response to drought (Alietal.2021).One way to improve drought tolerance in cultivated barley is by interspecific hybridization.Wild barley species,H.vulgaressp.spontaneum(orH.spontaneum),are diverse genetically (Lakewetal.2013),and relatively easy to cross with cultivated barley.Therefore,wild barley is frequently used for the introgression of novel alleles for drought tolerance and genotype enrichment using molecular markers (Lakewetal.2013),and studies have shown this can increase yield by 12-22% under drought conditions (Lietal.2007).

2.3.QTLs and molecular markers

The ‘new era’ of research on molecular markers and their association with quantitative trait loci (QTLs) emerged in the 1990s.The identified QTLs were associated with single or multiple genetic regions,based on the locations of other molecular markers on the genetic map.This was the initial step in our understanding or prediction of candidate genes located in an identified QTL region.While there is no doubt that QTL analysis and genetic maps with molecular markers have played an important role in crop plant improvement,QTLs still represent only a narrow (or wide) genetic region containing a target gene or genes,and not the genes themselves.Progress made with molecular markers has been applied to the fully sequenced genomes of various model plant species,such asArabidopsisthalianaandBrachypodiumdistachyon,as well as several important crops,and they are now publicly available as physical maps.In these genomes,almost all the genes are currently annotated,providing an opportunity to distinguish groups of genes in the QTL region and to develop new molecular markers for directly targeting genes based on the physical maps of the genomes (Cobbetal.2019).

Many QTLs that have been identified for barley are associated with drought tolerance (Comadranetal.2008;Kosováetal.2014).For example,QTLs associated with phenological traits were identified in a classical segregating population betweencv.Steptoe andcv.Morex,where a small number of genes had a large phenotypic effect (Kandemiretal.2000).Such examples in barley demonstrate that QTL identification is very closely related to the density of markers on the genetic maps,whereby more molecular markers in a genetic map improve its resolution and value.

However,each QTL in barley covers about 10 cM on average,which can contain several hundred genes on the corresponding physical map (Bernardo 2004).For example,based on meta-analyses of QTL in barley under drought,two meta-QTLs were identified (MQTL3H.4 and MQTL6H.2) but they covered genetic regions with 125 and 380 genes,respectively (Zhangetal.2017).A similar situation was found in a study of QTL localization for GY in barley introgression lines from a cross between cultivated and wild barley (Honsdorfetal.2017).The authors reported one unique QTL in chromosome 2H using the barley 9K iSelect genotyping platform,but this genetic region was so large (33.9-62.7 cM) that it covered hundreds or even thousands of genes.Therefore,QTL and candidate gene identification is only a useful starting point,and it is still far removed from the application of MAS for practical barley breeding.

2.4.Association studies and hybrid populations

Two different strategies are traditionally used for markertrait association analyses in barley and other crops.Genome-wide association study (GWAS) is based on the comparative analysis of well-established molecular markers on a genetic map in a large number of diverse accessions,preferably in different environments or over multiple years.GWAS is a very powerful method for studying the associations between genotyping and phenotyping,and is quite popular among researchers working with various crops,including barley (Chopra 2014;Abed and Belzile 2019).

In contrast,the second method is more traditional and requires a hybrid population with a relatively small number of progeny,like doubled haploids (DH),recombinant inbreed lines (RIL) or near-isogenic lines (NIL).Mutants and backcrossed lines with mutated alleles represent an additional option,in cases where a mutant hybrid population can be established and successfully used(Drukaetal.2011).Mapping populations based on advanced breeding lines,or from developed generations,are also very suitable for the molecular-phenotypic analyses when the resolution of the genetic maps is good enough with a sufficient number of established molecular markers.However,the success of molecular markers designed and developed for one population does not guarantee similar success in another unrelated population(Cobbetal.2019).

3.Assessment of molecular markers and candidate genes in barley plants under drought stress

All of the cases examined in this study were assessed for the availability of marker sequences,their locations on genetic maps,and their correspondence to barley physical maps.Positions of candidate genes and molecular markers were identified in the barley genome sequence ofcv.Morex,viathe Gramene website (see above).The retrieved sequences of molecular markers are presented in Appendix A,and descriptions of the genes and genetic fragments with identified molecular markers are given in Appendix B.These appendices show the complexity and diversity of markers and candidate genes that are available.The results include various barley genetic backgrounds,growing conditions,and stress treatments,with molecular markers which were identifiedviaboth manual and high-throughput assays,yielding results varying in integrity.

We focused on agronomic traits related to both GY and yield components.Drought can cause reductions in spikes per plant,grain per spike,and 1,000-grain weight (TGW),which all affect final GY (Alietal.2021).However,some experiments with germinated barley seeds,seedlings and very young plants have also been included for comparison with results from mature postharvest plant analyses.While it may not always be appropriate to compare physiological and biochemical responses of seedlings with mature plants,QTLs and possible candidate gene identification at early stages of plant development may be important for understanding the mechanisms of plant responses to drought.However,such responses may be less useful for MAS breeding of barley genotypes for improved drought tolerance and better GY.

3.1.Germination stage

During seed germination and the early stages of seedling development,26 QTLs representing dehydration-related traits were identified using GWAS from 218 spring barley accessions exposed to 20% polyethylene glycol (PEG).PEG treatment represents controlled dehydration and is a good drought simulation method,but these results cannot be fully assessed until the plants are grown to maturity and harvested (Table 1) (Thabetetal.2018).

3.2.Early seedling developmental stage (7-20-dayold)

In experiments with 113 barley cultivars,7-day-oldseedlings in pots were exposed to drought stress by withdrawing watering and keeping the soil water capacity at 20% for 4 wk (Wehneretal.2015).The authors reported around 47 QTLs,353 significantly associated SNPs,and 16 protein coding genes associated with drought stress.The most important of these were located in chromosomes 2H and 5H.The first SNP marker,SCRI_RS_239231,was reportedly located at 49.2 cM on the genetic map of chromosome 2H,and this marker was strongly associated with leaf senescence and chlorophyll loss during early drought (Wehneretal.2015).In our analysis using the annotated genome of barleycv.Morex,V3,this SNP marker was present between two very closely located genes,2HG0119030and2HG0119040.These genes encode a Cyclic nucleotide-dependent protein kinase (PTHR24353) and a SAC3 family protein B(PTHR12436),and both genes are estimated to be closest to the SNP marker.Sucrose synthase 4,SUS4,was reported to be associated with the SNP markerSCRI_RS_239231(Wehneretal.2015),but our search revealed that the closestSUS3gene,2HG0124890,was located more than 32 Mb away on the barley physical map and there were around 580 genes in the genetic interval.More studies are required to clarify which candidate gene was associated with the SNP marker in chromosome 2H and linked to leaf senescence (Table 1;Appendix B).

Table 1 Summary of the most important molecular markers,quantitative trait loci (QTLs) and genes,and their assessments in the barley genome,cv.Morex,and suitability for marker-assisted selection (MAS) targeting grain yield related traits in barley plants under drought and drought-simulated conditions (in order of ontogeny stages)

In another experiment,three SNP markers were identified as co-located in a 44 cM region on the genetic map of chromosome 5H (Wehneretal.2015,2016).Two of these markers,SCRI_RS_102075andBOPA1_ABC08327-1-1-353,were indeed located in the same gene5HG0438740and their annotated function was a complete match with that published (Wehneretal.2015,2016),i.e.,‘abscisic acid (ABA)-activated serine/threonine protein kinase,SAPK9(PTHR24343)’.This gene is involved in the drought stress responses in other plants (Xingetal.2022),and this supports the authors’conclusion that drought tolerance in barley breeding can be improved using MAS (Wehneretal.2015,2016).

The third SNP marker identified in those studies,BOPA1_9766-787,was located in the5HG0440850gene in the same chromosome,5H.However,it is not feasible that all the three markers could be co-located at the same position as published,at 44 cM on the genetic map.The distance on the physical map of barleycv.Morex between genes5HG0438740and5HG0440850is very large,and contains about 210 genes,covering a genetic fragment of more than 16 Mb (Table 1;Appendix B).Additionally,the annotation of the candidate gene5HG0440850with the located SNP markerBOPA1_9766-787was recorded as a purple acid phosphatase-related gene (PTHR45778),which is very different from the published nucleotide pyrophosphatase/phosphodiesterase (Wehneretal.2015,2016).Further study is required to clarify this issue.

In another experiment,10-day-old seedlings of an accession of wild barley,Hordeumspontaneum,were grown in hydroponics (Suprunovaetal.2007).Dehydration stress (3 and 12 h) was applied by removing the growth solution.TheHsdr4gene (H.spontaneumdehydration-responsive) was identified in chromosome 3HL between SSR markersEBmac541andEBmag705.This fragment on the genetic map spanned 26.4 cM,which is very large (Suprunovaetal.2007),and it covers 4,980 genes (3HG031840-3HG323320) based on the sequence of barleycv.Morex.However,using a different approach,the authors applied a method of transcriptderived fragments (TDFs),and only 11 TDFs from about 1,100 in total were identified as differentially expressed under rapid dehydration.Finally,only oneHsdr4gene was reported as the most suitable candidate for the dehydration-responsive gene,which showed homology to theRho-GTPase gene in rice (Suprunovaetal.2007).Based on this information,the corresponding gene in thecv.Morex genome can be identified as3HG0319800(Table 1;Appendices A and B).This candidate gene is related to osmotic adaptation in barley,but any potential effect on GY in adult plants remains unknown.

Seedlings at a similar stage (13-14-day-old) were studied in containers where watering was reduced to just 10-11% of the soil water capacity for drought conditions(Gudysetal.2018;Makhtoumetal.2022a).Nine and 64 QTLs were identified under this stress,respectively,associated with the control of various physiological traits,such as stomatal number,root weight,leaf weight,and leaf number,under drought.The authors used a variety of molecular markers,but we were unable to find the sequences of the markers to identify the closest candidate genes.Additionally,in some cases,the closest molecular markers were located as far away as 2.23 or 3.04 cM from the identified QTLs (Makhtoumetal.2022a).Such large genetic intervals can include tens or potentially a hundred genes.Therefore,more research is required to complete that study and to find the functional genes that are associated with these molecular markers.

In an experiment with 18-day-old barley seedlings from a segregating population that were grown in pots with watering withdrawn (32% soil water capacity at day 7),many molecular markers were identified as being associated with various physiological characteristics.These included the efficiency of photosystem II (PSII),gas exchange and some recognized drought stress index indicators (Rapaczetal.2010;Fiustetal.2015).The DArT-based markerbpb-0994(2H,107.8 cM),and two SSR markers,scssr02503(5H,42.8 cM) andBmag500(6H,31.7 cM),were strongly associated with these physiological parameters.The SSR markerscssr02503was associated with a QTL containing a possible candidate gene that was similar to auxinresponsive protein IAA30-like (Fiustetal.2015).Our analysis of the barley genomecv.Morex,chromosome 5H,revealed that the amplicon of thescssr02503marker has a perfect match to a sequence within the 3´-UTR of gene5HG0433470.This gene indeed encodes auxin-responsive protein IAA17 (Appendix B),which is functionally very active under conditions of various abiotic stresses (Shietal.2020).The markerbpb-0994was found to be located in chromosome 2H of the barley genomecv.Morex,at gene2HG0190740,and encodes a zinc ion-binding protein.The last marker,Bmag500,was located in an intergenic region in chromosome 6H with the closest gene being6HG0545620incv.Morex,which was annotated as a heat-stress protein 70 (HSP70).The authors noted that their chosen molecular markers were effective in MAS under drought (Fiustetal.2015).However,a functional analysis of these and nearby candidate genes was not provided,so this assertion is difficult to assess.

3.3.Middle-stage plant development (1-2-monthold)

Plants of a DH population originating from a barley hybrid (TX9425×Franklin) were grown in pots with soil and were treated with drought for one month following the onset of the tillering stage (Fanetal.2015).One major QTL for drought tolerance was identified in chromosome 5H (133.7 cM) based on the DArT markerbPb-3241.The authors described 448 genes located in the genetic interval of the QTL,and a candidate geneMLOC_18300.1,encoding 9-cis-epoxycarotenoid dioxygenase 2,was identified as the most promising.Based on our search of thecv.Morex sequence,the gene5HG0508150was identical as reported by Fanetal.(2015),and the DArT markerbPb-3241was located in the genetic fragment containing 274 genes from the candidate location.This comprises a very large genetic interval,and the selection of a single gene encoding “an important enzyme during ABA synthesis under drought stress” (Fanetal.2015) may be considered ambitious since it was not accompanied by any functional analysis.

Plant growth characteristics of selected breeding lines at the mid-stage of development (1-2-monthold) were evaluated using a conveyer imaging system,where drought was simulated by withholding water(Phametal.2019).Under drought,the authors identified developmental genesHvCEN(Centroradialis) on chromosome 2H and theVrn-H2vernalization gene,which was co-localized withBFL(BarleyFloricaula/Leafy).These genes were associated with all the developmental traits examined,including dry weight,tiller number,plant height,growth rates and water use efficiency (Phametal.2019).However,no assessments of GY or yield components were presented in that study.

3.4.Reproductive-anthesis stage of plant development

Barley plants of 121 accessions exposed to drought stress(35% field capacity) in the early reproductive stage and analyzedviaGWAS showed more QTLs and candidate genes related to yield component traits,including spike length,spikelets per spike,grains per spike,and TGW(Thabetetal.2020).The authors reported five genetic regions located in chromosomes 2H and 3H,with 12 SNPs significantly associated with these traits.Major effects on spikelet number and final grain number per spike under drought stress conditions were strongly associated with two SNP markers (SCRI_RS_166540andSCRI_RS_157347),identified on chromosome 2H.The authors reported that these two SNP were located in two candidate genes designated asHORVU2Hr1G091030and2Hr1G091170(Thabetetal.2020).These candidate genes correspond toHORVU.MOREX.r3.2HG0183720and2HG0183770,respectively,in thecv.Morex genomeviathe Gramene web-site,and they are annotated as RNA polymerase II C-terminal domain (CTD) phosphatase and Expansin-B3,respectively (Thabetetal.2020).Therefore,the candidate genes match perfectly with the molecular probes designed to target the SNP in this case(Table 1;Appendix B).

Two other SNP markers identified by Thabetetal.(2020),BOPA1_2391-566andSCRI_RS_177313,were significantly associated with number of spikelets per spike and TGW,respectively.They were mapped on chromosome 3H,and the authors identified the corresponding genesHORVU3Hr1G019590,MYB-domain protein 37,and3Hr1G098200,a homolog of a genomic scaffold of wheatcv.Chinese Spring on chromosome 3B,respectively,as the best candidate genes for these SNP markers.However,our analysis revealed that the most likely candidate genes for these two SNP markers were3HG0236620,a metalloprotease similar to the rice protein Os01g0191500,and3HG0313760,a F-box family protein 31/39,respectively.These are presented in Appendix B,and it is noteworthy that these genes were located far from the proposed genes (more than 40 Mb and 330 kb,respectively) on the physical map of barleycv.Morex(Thabetetal.2020).Additional study is required to obtain a more precise identification of the candidate genes.

At the reproductive stage,drought was applied to 108 plants of F8inbred lines from a hybrid (Badia×Kavir)(Makhtoumetal.2022b).Among many studied traits,a single QTL for spike weight was identified under drought conditions in a glasshouse.The closest SSREST marker to the QTL wasGBM1212,which is in chromosome 6H.We found the marker resides within the gene6HG0554560of the barley sequencecv.Morex,encoding the known transcription factor,bZIP,class 44.Possible candidate genes in this and other QTLs were not examined (Makhtoumetal.2022b),so this study is not yet suitable for MAS.

Pre-anthesis drought tolerance also was studied in a panel of 100 diverse barley accessions using plant imaging and GWAS analysis with a 9K SNP Illumina array(Dhanagondetal.2019).Several QTLs and marker-trait associations were identified,particularly for tiller number in chromosome 5H at 120.1 cM on the genetic map.Two well-characterised genes,Vrn-H1andHvPhyC,are located in this region,and they control plant responses to vernalization and changes in the response to the light spectrum,respectively (Dhanagondetal.2019).However,many other genes are also co-located in this QTL,so it is difficult to identify any candidate genes without further functional analyses.

At anthesis,plants of a DH hybrid population were subjected to severe drought at 20% soil field capacity in pots (Gousetal.2016).QTLs with DArT markers were linked to ‘stay-green’ traits.Using the barley genome ofcv.Morex,the markersbPb-6023andbPb-3703were found which are in chromosomes 6H and 7H,respectively,in the genes6HG0556600,encoding a Flavin-binding monooxygenase-like protein,and close to7HG0722350,encoding a UBX-domain containing protein.However,no further information about potential candidate genes was provided (Gousetal.2016).

3.5.Fully mature plants and the harvest stage

In Mediterranean climate conditions with terminal drought,the most confident QTLs specific to the genetic background were mapped on chromosomes 1H and 6H(near markercdo497) for TGW (Teulatetal.2001).In similar studies over four consecutive years in Syria and Lebanon,recombinant inbred lines (RILs) of a barley hybrid population (Tadmor×ER/Apm) were tested (von Korffetal.2008).Thirty-eight QTLs were found for 11 traits related to GY.The strongest effect for GY,with 17.6% of the genetic variance,was reported for QTL-6Hb with the RFLP markercdo497.The authors reported about 200 kg ha-1higher GY in a RIL with the paternal(ER/Apm) allele of thecdo497marker.Major effects on multiple traits including plant height,lodging,peduncle length and peduncle extrusion for GY in negative relations were shown for QTL-3Hb with the RFLP markerbcd1127(von Korffetal.2008).However,in a meta-QTL analysis together with other studies,the intervals for these two QTLs at 6H and 3H,designated as MQTL-6H.2 and-3H.4,were found to be very large and they contained 380 and 125 candidate genes,respectively (Zhangetal.2017).

In the experiments described above,all the drought or dehydration treatments started from a single time-point in plant development.In contrast,three stages of plant ontogeny were used in an analysis of barley plants from 94 DH lines originating from a (DOM×REC) mapping population,including (1) seed germination in 15%PEG-6000;(2) a 3-4-leaf seedling test in hydroponics with 15-18% PEG;and (3) adult plants in different growth conditions (Sziraetal.2011).Based on ESTSSR markers,the authors reported three QTLs and corresponding candidate genes that were responsive in all conditions.The successful EST markers were: (1)GBM1498,2H;(2)GBM1404,6H;and (3)GBM1359,7H,and the genes to which these markers were mapped(Sziraetal.2011).The corresponding genes in thecv.Morex genome were: (1) DREB2,ethylene-responsive transcription factor RAP2-1-related (2HG0204820);(2)nuclear inhibitor of protein phosphatase-1 (6HG0631630);and (3) PP2A,serine/threonine-protein phosphatase(7HG0716180) (Appendix B).Additive effects of these three markers were reported to be also associated with genotypes producing higher and lower GY,respectively.The authors hypothesised that the EST-SSR markers can be used in MAS for the improvement of drought tolerance in barley breeding lines (Sziraetal.2011).

A GWAS analysis with 223 barley accessions was conducted using 61 SNP,45 SSR and 710 DArT markers in a genetic map.The plants were grown in two field trials,with high rainfall (4.4 t ha-1) and low rainfall (1.4 t ha-1) until harvest (Baumetal.2003;Varshneyetal.2012).For GY in the dry field trial,the most significant associations were found with SNP markerGBS0469on chromosome 1H,133 cM,explaining around 3.8% of the phenotypic variation.This marker was located in the gene1HG0089460,1H: 503425301-503430730,named OS01G0663400-like protein,beta-site APP-cleaving enzyme,isoform A (Varshneyetal.2012).

In other studies,almost all QTLs for GY traits were associated with phenological traits and with genetic fragments showing strong co-location with genes known to control plant development (Tondellietal.2014).In Mediterranean environments,a DH mapping population from southern Europe was analysed with plants grown in ‘wet’ and ‘dry’ field trials.Grain yield under drought conditions was associated with three DArT markers and the corresponding genes were: (1)Hv347D22-HvFT3(Ch1H,76.6 cM) orHvFT3gene;(2)HvCen-Eps2(Ch2H,78.7 cM) theCentroradialisgene,HvCEN4,where theFT3andCEN4genes are both involved in time to flowering;and (3)HvBM5A-Vrn-H1(Ch5H,122.2 cM)the vernalization geneHvVrn-H1.The conclusion was that the majority of the identified QTLs for GY traits under drought were located exactly in the three genes above(Tondellietal.2014).

In central Chile,with a relatively similar Mediterranean climate,recombinant chromosome substitution lines(RCSLs) originating from a cross between barleycv.Harrington andH.spontaneum,Caeserea 26-24,were studied in both irrigated and rain-fed conditions (Inostrozaetal.2009;Moraetal.2016).For GY,two stable QTLs were detected in dry (rain-fed) conditions,with SNP markers2711-234and1923-265,both in chromosome 1H,at 139 and 140 cM,respectively,explaining 7 and 8.6% of the phenotypic variation.The positions of the2711-234and1923-265markers on the genetic map of chromosome 1 were slightly different at 100.69 and 101.45 cM,respectively (Closeetal.2009).In the results,the first marker,2711-234,was identified in the gene1HG0078320(cv.Morex),encoding a Dim1 family protein or Thioredoxin-like protein 4A (PTHR12052).The second marker,1923-265,was found in the gene1HG0079230,encoding a PDCD5-like protein,programmed cell death protein 5 (PTHR10840).These SNP markers,based on EST probes,are presented in Appendix A,with full sequences of the corresponding genes in the genome ofcv.Morex (Appendix B).

Three DH barley populations were tested under drought conditions in Australia,and several QTLs were identified for GY,explaining up to 25% of the phenotypic variation (Obsaetal.2017).However,the authors reported that major developmental genes,includingPpd-H1,Vrn-H1,Vrn-H2andVrn-H3,did not affect yield performance.One of the major QTL for GY was identified in chromosome 2H,between GBS markersTP23249andTP45335.However,this genetic region on the physical map of barleycv.Morex,v.1 (515,418,972-515,910,655)covered 22 genes,from2HG0173530to2HG0173750.Further study is required to identify and functionally analyse the possible candidate genes (Obsaetal.2017).

A diverse panel of 192 genotypes was studied for root traits under drought using genome-wide epistasis analysis (Oyigaetal.2020).The genetic intervals of the QTLs contained three important candidate genes: (1)ZIFL2,zinc induced facilitator-like 2;(2)MATE,MATE efflux family protein;and (3)PPIB,peptidyl-prolylcistrans isomerase B.These genes are related to the nodal root response and adaptation to drought,and the authors concluded that the candidate genes can be used for MAS targeting root-related traits (Oyigaetal.2020).

At post-anthesis,drought simulationviafoliar application of the desiccant KI was used on a diverse collection of 183 barleys (Tarawnehetal.2020).Based on GWAS,the authors identified 28 markertrait associations identical to those in QTLs with perfect localization on the physical map,and 10 candidate genes were reported.Some of the genes were closely colocated or repeated,and so this list was reduced to the seven most suitable genes,which mainly control grain weight and grain number per spike.The identities of the molecular markers and genes in the barley sequence genome ofcv.Morex are presented in Appendices A and B,and in Table 1.However,the authors emphasized that functional validation will be required to confirm the identified genes and their role in plant responses to drought (Tarawnehetal.2020).

In an analysis of 148 barley cultivars over two years,plants were grown in field trials in irrigated (control) and drought conditions (40% of filed capacity,FC) for postharvest analysis (Jabbarietal.2022).Numerous AFLP markers were found to be associated with GY in drought conditions,such asE38M50-242.In the absence of any sequence or marker localization data,the candidate genes cannot be identified for further MAS.Nevertheless,the authors stated that their identified markers could potentially be used in MAS for barley drought tolerance(Jabbarietal.2022),but this appears to be overly optimistic.

In a GWAS analysis of 426 spring barley breeding lines of diverse origins,Elbasyonietal.(2022) applied drought over the entire lifespan from the early seedling stage until harvest.For GY in drought,eight markers were identified as strongly associated with candidate genes,and one SNP marker,11_20012,in chromosome 4H showed the highest probability for association.The authors identified the candidate gene asHORVU4Hr1G010700with potential functional annotation (based on rice homology)as protein 50S ribosomal L12-2,a chloroplast precursor(Elbasyonietal.2022).In thecv.Morex genome,4HG0339990is still listed as having a transcript with unknown function.The next three genes in the distal direction (4HG0339960-4HG0339980) encode very short polypeptides that are Serine-type endopeptidase inhibitors,while the three genes in a proximal direction ending with4HG0340020were identified as MYB DNA-binding 3-domain-containing protein,similar to Os02g0813200=LOC_Os02g56830 (MYB-like,SANT family with transposon CACTA) in plants and other organisms (Ambawatetal.2013),which has variable functions including cuticular wax biosynthesis and drought tolerance (Liuetal.2015).On the physical map ofcv.Morex,the 16 kb region between4HG0339990and4HG0340020contains only two small genes.Therefore,4HG0340020can be considered as an alternative to the published candidate gene4HG0339990(Elbasyonietal.2022).In the absence of any functional analyses of the candidate genes,both the published and other alternative candidate genes remain speculative.

4.Six research strategies: From molecular markers and genes to MAS

Molecular markers linked to specific QTLs,or developed with association to candidate genes,can be employed by barley breeders to improve the selection of desirable genotypes and accelerate breeding programs (Baumetal.2007;Kosováetal.2014;Cobbetal.2019).This central aim for the practical use of molecular markers is illustrated by the six strategies of MAS development,showing their similarities and differences (Fig.1).Among the many types of molecular markers,their nature,design and working principles are very diverse but they all remain genomic ‘tags’ (Henry 2013;Poczaietal.2013;Grover and Sharma 2016;Chengetal.2017).

Fig.1 Schematic representation of six strategies for the application of molecular markers to marker-assisted selection (MAS),starting from SSR-type (A) and SNP-type of molecular markers (B).The importance of various QTLs depending on stage of plant development (C);and the requirement for verification of gene expression (D) are shown.The RNA-seq strategy (E) does not specify any molecular markers and has no direct outcome to MAS,so is indicated as ‘under question’.The gene identification strategy (F) shows an alternative procedural order for the development of molecular markers for MAS.

4.1.Strategy A -SSR or similar types of molecular markers

Simple sequence repeats (SSR) have taken ‘central stage’as the preferred molecular markers for QTL mapping over the last 15-20 years (Parketal.2009;Vieiraetal.2016).SSR represent segments with a tandem repeat motif of 1-6 nucleotides,and often they are used because there is insufficient variation in the remainder of the genome after many generations of breeding.Requiring only simple PCR preparation,and highly polymorphic in nature,SSR markers continue to represent favourable and non-expensive molecular tools for genotyping and the determination of genetic polymorphisms.However,the first major limitation in applying SSR markers relates to their abundance and often very dispersed distributions across plant genomes,barley included.Secondly,SSR markers are mostly located in intergenic regions where reduced selective pressures permit greater variation.Only in rare cases,SSR were found in the introns or promoter regions of genes (Lietal.2004).Therefore,the use of SSR markers alone is often impractical for locating candidate genes,even in relatively narrow QTL genetic regions.

To illustrate this point,three SSR markers were selected from various publications associated with GY under drought conditions in barley (Sallametal.2019),wild barley,H.spontaneumL.(Abou-Elwafa 2016),and hybrid breeding lines from crosses between cultivated and wild barley (Kalladanetal.2013).In the first study,the most promising SSR marker,gwm292_1849in chromosome 7H,was associated with increased TGW,explaining 11% of the phenotypic variation (Sallametal.2019).The SSR markerXgmw292from bread wheat had 10 matches in the barley genome,with the greatest similarity reportedly mapped to the distal region of chromosome 7H.In the barley genome ofcv.Morex,our analysis of theXgmw292marker revealed the intergenic region as 7H: 600,636,124-600,636,140,between two flanking genes,7HG0738740and7HG0738730.These genes encode relatively short polypeptides (172 and 242 aa),both with unknown functions.The authors made an optimistic conclusion about “choosing highly promising markers for enhancing drought tolerance in barley”(Sallametal.2019).However,this work must be verified with more precise identification of the possible candidate genes within the region ofgwm292_1849.

Similar results were found in two other examples.SSR markerEBmac0788,mapped on chromosome 4H(97.67 cM),was associated with TGW and explained the highest amount of phenotypic variation at 27.8%(Abou-Elwafa 2016),while SSR markerBmac90on chromosome 1H (78.5 cM) was associated with GY and seed size (Kalladanetal.2013).The authors did not find genetic regions or possible candidate genes.Our analyses of these SSR markers are presented as markers No.21 and 2,respectively (Appendix B),with primers and PCR amplicons.Finally,for SSR markerEBmac0788,the corresponding intergenic region in chromosome 4H was identified between genes4HG0407970and4HG0407980 incv.Morex (No.21;Appendix B).The corresponding intergenic region in chromosome 1H was identified between genes1HG0049280and1HG0049290(No.2;Appendix B).These genes were classifiedviathe presence of domains,but their encoded proteins remain uncharacterized.Therefore,bothEBmac0788andBmac90markers (Kalladanetal.2013;Abou-Elwafa 2016) need further investigation.

The simple conclusion to draw from these three examples is that since SSR markers are mostly located in intergenic regions,finding appropriate candidate genes involved in the associations is not always possible.While they could be among the closest flanking genes,there is no way to judge the genetic distance that the genes may be from the SSR markers.Therefore,SSR markers can only be useful when they are adequately validated through identification and verification.However,SSR markers can still be helpful for the development of more advanced markers,as well as for reducing the number of selected linesviapreliminary MAS.

4.2.Strategy B -SNP and similar markers,relationship with haplotypes

In contrast to SSR,single nucleotide polymorphism (SNP)markers may be located in any part of the genome,within either candidate genes or intergenic regions (Mammadovetal.2012;Huqetal.2016;Rasheedetal.2017).In intergenic regions,the situation is similar to that described above for the SSR markers.Flanking genes in both directions on the physical map may be considered as possible candidate genes,and a SNP found inside an open reading frame (ORF) is likely to be related to that particular candidate gene.The value and importance of synonymous SNPs (causing no amino acid change) and non-synonymous SNPs (causing a change) in coding regions are obviously very different and discussed thoroughly in many papers,including in barley (Liao and Lee 2010).The regulatory role of SNPs in promoter regions of genes in plants also has been discussed often(Czajkowskaetal.2019).

SNP markers are used in diverse plant species,including barley,and they have become a popular topic among researchers (Boopathi 2020).Additionally,other genotyping methods are based on SNPs,from cleaved amplified polymorphic sequences (CAPS) (Shavrukov 2014,2016) to microarray technologies (Talamèetal.2007;Bedadaetal.2014),and they also have been successfully applied in barley MAS (Haydenetal.2010).

Groups of SNPs that are inherited together as one‘genetic block’ or fragment are often reported.These socalled ‘haplotypes’ are also important for genotyping,SNP marker development and applications in MAS (Lorenzetal.2010).It is also noteworthy that small insertions or deletions (InDel) are sometimes part of these haplotypes in addition to SNPs.

In summary,since the accuracy of marker ‘prediction’is strongly dependent on the distance to a potential candidate gene on the genetic or physical map,markers like SSR that are mapped in intergenic regions provide limited options for the identification of candidate genes,whereas SNP markers located within ORFs or in promoter regions are likely to be associated with the candidate gene.

4.3.Strategy C -During ontogeny,plants react differently in response to drought

Many studies have investigated different QTLs and candidate genes.One of the reasons for such variability is the wide range of plant developmental stages among the plant materials used in experiments,ranging from germinating seeds and seedlings,to mature plants and harvested produce.Therefore,it seems obvious that different genes would be involved in the reactions to drought throughout plant development.

A simple way to support this point is to compare barley development during the vegetative and reproductive stages.The transition to tillering and flowering is controlled by a large group of genes,including the vernalization response,photoperiod,and flowering time.Molecular markers can be either related or unrelated to such developmental genes.For example,in a Spanish breeding program,theVrnH1andPpdH1genes showed strong co-segregation with all GY-related traits for barley breeding from the F2 to F8 generations of lines grown in dry environments,where DArT-seq markers were employed for molecular analysis (Igartuaetal.2015).In many other cases described above,the identified QTLs and candidate genes were not associated with developmental genes in barley.It is important to select the most appropriate stage of barley plant development,preferably the one that is the most typically affected by drought in the specific environment.However,it is still likely that the results will differ from those obtained in experiments with plants at other ontogenetic stages.

4.4.Strategy D -Verification and confirmation of gene expression ‘from molecular markers to MAS’

Prospective candidate genes derived to explain any phenotype or adaptive response should be confirmed independently.Verification of candidate gene expression represents the simplest way to confirm that the tested candidate gene is indeed stress-responsive,such as under drought conditions.

However,relatively few papers in this field provide any verification of the candidate genesviagene expression.In two examples,theHsdr4gene (Hordeumspontaneumdehydration-responsive) was identified and tested for gene expression in 10-day-old seedlings (Suprunovaetal.2007);while two other genes,HVA1,late embryogenesis abundant,andSRG6,stress-responsive gene 6,were identified and confirmed by gene expression using qPCR in malt-and feed-types of barley (Rapaczetal.2010).In another experiment,one gene was confirmed as highly expressed under drought among the five genes studied;and this gene,HvNCED,9-cis-epoxycarotenoid dioxygenase,is involved in ABA biosynthesis and was confirmed to have a drought-responsive expression profile(Harb and Samarah 2015).In another study with 156 barley genotypes and 14 identified candidate genes,two genes were confirmed to have strong expression levels in response to drought:TRIUR3,abscisic acid-inducible protein kinase,andAVP1,Arabidopsis-like vacuolar pyrophosphatase.These genes were co-located in a QTL mapped at 45 cM on chromosome 5H (Wehneretal.2016).

The ‘in-silico’ analysis of available RNA-seq data can be used for the confirmation of expression as a preliminary estimation,particularly when gene expression profiles are shown for unstressed barleycv.Morex(Tarawnehetal.2020).In such cases,the genotypes should be checked for gene expression under drought conditions.

4.5.Strategy E -RNA-seq microarray technology

The advent of modern RNA-seq microarray technology has opened up new possibilities as thousands of genes can be studied simultaneously,such as in comparisons of barley genotypes with contrasting responses to drought.This strategy provides an accurate and comprehensive overview of differentially expressed genes.However,further research is required to analyse all these genes and identify the most important genes for drought tolerance and yield production (Talamèetal.2007;Bedadaetal.2014).

The following RNA-seq microarray reports represent typical examples of the outcomes highlighting anywhere from a few to several hundred candidate genes.Firstly,263 genes were differentially expressed at the reproductive stage among three barley genotypes,cv.Martin,cv.Moroc9-75,andH.spontaneumline 41-1,HS41-1 (Guoetal.2009).The authors selected 12 genes for further validationviaqPCR expression analysis,but only a correlation plot was presented,with a moderate correlation value (R2=0.85) between the microarray and real-time results (Guoetal.2009).

Another report examined four wild accessions ofH.spontaneum,two tolerant and two sensitive to drought (Hübneretal.2015).In total,2,768 genes were differentially expressed in all four,while 602 and 2,166 genes were specific for the tolerant and sensitive accessions,respectively.Four candidate genes for drought-tolerance were selected with different alleles.Two genes were associated with drought,dehydrogenase/reductase (AK362742),and vacuolar protein(AK368692).The third gene,GPI mannosyltransferase(MLOC_67950.1),was related to pollen viability,and the fourth (AK370720) has no known function (Hübneretal.2015).

In another example,a Spanish landrace-derived line(SBCC07) andcv.Scarlett were studied (Cantalapiedraetal.2017).The number of genes in thedenovoassemblies were 112,923 in SBCC073 and 123,582 in Scarlett under drought.However,only five genes were identified as being differentially expressed in the leaves of the two barley genotypes,including: (1)polyamine oxidase,involved in spermine and spermidine degradation;(2) down-regulation of chlorophyll apoprotein from photosystem II;and three genes encoding proteins,(3) ABA/WDS (abscisic acid/water deficit stress)induced,(4) ribonuclease T2,and (5) calcineurin-like phosphoesterase (Cantalapiedraetal.2017).

A comprehensive comparison of original data and other similar results using microarrays in barley was presented.The authors showed 23 differentially expressed genes that were common in at least three published papers(Cantalapiedraetal.2017);but after filtering the results to show only drought,and excluding young seedling stages,seven genes were common to two publications.Only a single gene was common to all three publications,namelyHSP70,heat shock 70 kDa protein 1/8.However,after sequencing,it was impossible to determine which gene ofHSP70was analysed among the 20 genes in this family in the barley genome (Landietal.2019).Additionally,the expression patterns of the selected genes were analysed using qPCR but only low to moderate correlations(R2=0.32-0.79) were shown between the microarray and qPCR analyses (Cantalapiedraetal.2017).

The results of the last two examples of this strategy are inconclusive.Barleycv.Golden Promise (GP)andcv.Otis were exposed to drought in the first study(Mahalingametal.2022).About 4,650 and 3,905 genes were differentially expressed genes in flag leaves of GP and Otis,respectively,while in spikes there were 2,467 and 1,576 genes in plants under drought stress.With so many identified genes,it was unrealistic to make any conclusions for MAS,and checking for potential candidate genes would require a much smaller number of selected genes (Mahalingametal.2022).In contrast,barleycv.Tadmor and exotic Tibetan wild barley XZ5 (H.vulgaressp.agriocrithon) were treated with PEG-induced dehydration (Zhangetal.2019;Qiuetal.2023).From 15,361 significantly affected genes in total,26 genes associated with dehydration were identified in the XZ5 genotype.Two novel genes,zinc-induced facilitator-like 2(HvZIFL2) and peroxidase 11 (HvPOD11),were selected as the most responsive to water limitation,and confirmed with gene expression data.These results were verified among 140 barley accessions (Qiuetal.2023).Based on the haplotypes of theHvZIFL2andHvPOD11genes,molecular markers were suggested for further application in MAS to improve drought tolerance in barley.

4.6.Strategy F -Reverse genetics: from identified genes to phenotypes,molecular markers and finally to MAS

Strategies A-D above are based on a ‘Forward Genetics’approach,where desired phenotypes (for example,better yield under drought) are studied first,with a subsequent analysis of QTL to find the genes that may be responsible(Cobbetal.2019).In contrast,‘Reverse Genetics’ begins with a gene already demonstrated to be important for the phenotype.Genes with known functions may be chosen,such as from microarray RNA-seq data (see above Strategy E).Strategy F demonstrates the latter procedure to find candidate genes,develop molecular markers,and test for MAS in barley for improved drought tolerance.

Many publications report on mostly well-known,drought-responsive genes,often in barley genotypes with contrasting stress responses and supported by gene expression profiles (Baumetal.2007;Lietal.2007;Güreletal.2016;Maroketal.2021).Although valuable,one limitation of such reports is the lack of molecular markers which could then be tested for MAS.As an example,two out of 16 genes,HvZIP1,encoding a bZIP transcription factor,andHsdr4,a homolog of the riceRho-GTPase gene,showed associations with a better response to drought in seedlings of nine barley cultivars and breeding lines (de Mezeretal.2014).While this was a good study of gene expression,it has not been extended and cannot be implemented in MAS without the development of specific molecular markers.

Similarly,in a report describing theP5CS1gene (delta-1-pyrroline-5-carboxylate synthase gene 1),which is involved in proline biosynthesis under drought stress (Xiaetal.2017),287 barley accessions were sequenced using EcoTILLING (a modification of Targeting Induced Local Lesions in Genomes).The results showed 16 SNPs and 25 InDels in the gene,with 13 distinct haplotypes.Five haplotypes were associated with drought tolerance traits,and two haplotypes (HvP5CS1_H1andHvP5CS1_H4) showed improved drought tolerance and better GY.Despite the absence of validation,this work provides very precise information on the SNPs,so it represents a good opportunity for completion through the development of molecular markers for theHvP5CS1gene and their application in MAS (Xiaetal.2017).

In our own work,SNPs were identified in two genes,HvSAP8andHvSAP16(stress-associated proteins,SAP),encoding zinc-finger transcription factors with A20/AN1 and AN1/C2H2 domains,respectively (Baidyussenetal.2021).Expression of these genes in response to drought showed different patterns in the lines of two mapping populations,G×B (Granal×Baisheshek) and N×A(Natali×Auksiniai-2).The proposed SNP markers were finally,successfully applied in MAS (Baidyussenetal.2021).Interestingly,similar genes from the same family,i.e.,zinc finger AN1 and knuckle domain proteins,were identified in a meta-data analysis of drought tolerance in rice with homologous searching in other cereals,including barley (Swamyetal.2011).

5.Conclusion and future prospective

In conclusion,we address the following question: Despite numerous reports of molecular markers and candidate genes identified as associated with drought tolerance in barley,why are there so few attempts to translate this into MAS for a breeding context?

Traditional methods using individual samples and gel-based evaluation are still available.Some reports describe a mixture of different strategies,and the choice of which strategy and breeding program to use will depend on the specific circumstances (Cobbetal.2019).Nevertheless,more modern technologies of plant genotyping are being reported,leading to the identification of many candidate genes,the development of greater numbers of molecular markers and testing for MAS(Rasheedetal.2017;Kumaretal.2020).The efficiency of molecular work has improved with widespread use of the GWAS method (Riazetal.2021),proteomics (Ghataketal.2017),and robotic and high-throughput technology platforms (Closeetal.2009).

MAS is a complementary method for use in conjunction with conventional breeding that has been successfully applied in different crops around the world (Lietal.2007;Ben-Arietal.2012;Boopathi 2020).Why then,are we still seeing such slow progress in the translation and application of molecular markers? As shown above,many QTLs and possible candidate genes have been suggested for the improvement of drought tolerance through barley MAS.To date,however,few QTLs and candidate genes have been fully realized as efficient molecular markers used for MAS in the breeding of barley (Ben-Arietal.2012) or other crops (Cobbetal.2019).

There may be several reasons for this situation.It is well known that the response of plants to drought is a quantitative trait,involving multiple genes with complicated interactions,at different developmental stages,and involving various systems and strategies,so multiple QTLs are expected to contribute to the phenotypic variability (Mohanetal.1997;Chopra 2014;Güreletal.2016;Riazetal.2021;Elakhdaretal.2022;Fatemietal.2022).Because of this complexity,the limited use of selected markers likely reflects the limitation of traditional MAS for highly complex traits such as drought,and this only will be solved with wider ranges of molecular markers,for example,viamicroarrays (Lande and Thompson 1990;Sallametal.2015).

The different stages of plant development,together with the intensity,timing and duration of drought or simulation of stress conditions,will collectively cause different and variable responses in plants,and varying expression of the stress-responsive genes (Tardieu 2012;Elakhdaretal.2022).Only 43 molecular markers and corresponding candidate genes were selected for this review from among the wide range of possible genes in the literature (Appendices A and B),and it is difficult to find co-incidence among the identified genes.They are widely different among the publications.From our point of view,this dilemma should not be taken as any proof of error,since the conditions in each experiment generally differ from the others.So,it is unrealistic to expect great similarities among them.

Our assessment of the results,as presented in Table 1 and described above,indicated that many QTLs and molecular markers were identified based on the available genetic maps.We suggest the next steps should be to transfer this information to the physical map,for example using barleycv.Morex.Most of our attempts at this were successful,and the corresponding genes or genetic fragments of thecv.Morex genome were identified.Finding either confirmation or newly identified genes makes the final step toward the development of suitable molecular markers closer.However,in some cases,this milestone was not easy to achieve in the absence of any sequence information for the molecular markers used.The published genetic maps are useless if only the names of molecular markers are provided without their respective sequences.Many marker sequences were retrieved from the GrainGene database (www.wheat.pw.usda.gov) and are presented in Appendix A.We also acknowledge the other researchers who have provided full sequences of their molecular markers in the supplementary or additional files for their publications,and we encourage other authors to do the same.Our Appendices A and B provide the necessary and important information about the published molecular markers,genes and genetic regions based on the genome of barleycv.Morex.Additionally,many authors agree that functional analysis of the candidate genes is required to confirm changes in gene expression in response to drought stress.Based on the verification of candidate gene expression,specific molecular markers may be developed for MAS,targeting GY-related traits in barley plants grown under drought conditions.

Finally,in conclusion,we provide one example that demonstrates the successful MAS application for GY improvement in barley,found in near isogenic lines (NILs)developed from the hybrid (Baronesse×Harrington)(Schmiereretal.2004).From 186 BC2F1progenies,the authors selected 17 NILs with predicted genetic fragments on chromosomes 2H and 3H.Only one NIL,00-170,had the confirmed presence of the predicted genetic fragment in chromosome 3H,and it generated significantly higher grain yields in 22 environments.Despite the absence of drought tests in this study,based on the results of MAS and using flanking molecular markersMWG571AandMWG961,the NIL line 00-170 was selected and confirmed for genotyping and phenotyping (Schmiereretal.2004).

This study was carried out 20 years ago,but even now it represents excellent and elegant research with accurate and precise identification of the genetic regions with targeted chromosome fragments.In our modern time,this example should be considered for the further identification of similar candidate genes,development of effective molecular markers and their successful application in MAS to improve drought tolerance in crops such as barley.

Acknowledgements

This study was supported by Bolashak International Fellowships,Center for International Programs,Ministry of Education and Science,Kazakhstan;and Research Projects BR10764991 and BR10765000 supported by the Ministry of Agriculture,Kazakhstan and AP14869777 supported by the Ministry of Education and Science,Kazakhstan.We want to thank staff and students from our affiliations for their support in this research and help with critical comments to the manuscript.

Declaration of competing interest

The authors declare that they have no conflict of interest

Appendicesassociated with this paper are available on https://doi.org/10.1016/j.jia.2023.06.012