Climate change is one of the major challenges of the 21^st century. The Intergovernmental Panel on Climate Change (IPCC, 2023) predicts an increase in average global temperatures between 1.4 and 4.4°C by the end of this century. Increased global temperatures are predicted to have vast flow-on effects on global climate and local weather systems leading to more extreme and unpredictable temperatures, as well as increased extreme rainfall, drought, storm and fire events (IPCC, 2023). Many of these impacts are already being experienced around the world, with NASA declaring July 2023 the hottest month on record fuelling heat waves, wildfires, and floods across the Northern Hemisphere. These changes will affect every tree crop industry and region to varying degrees. Temperate tree crops are at particular risk due to their high reliance on seasonal variations in temperature for processes such as bud dormancy and flowering (Campoy et al., 2011), and pollen development and fruit set (Irenaeus and Mitra, 2014), which all in turn impact yield. Subtropical tree crops are also not immune to the effects of climate change with many reliant on cold winter temperatures to induce flowering (Wilkie et al., 2008), and temperature also significantly impacting pollen and fruit development (Dinesh and Reddy, 2012).

Another challenge facing humanity this century is increasing population size and demand for food. Taagepera and Nemčok (2023) predict the global population to increase to a peak of 11.2 ± 1.5 billion by the end of this century. Population increases during the 20^th century were offset by the Green Revolution, a doubling in production of high-yielding varieties of several cereal crops (Khush, 1999). However, population growth since the 1990s has outstripped the rate of growth in food production requiring new strategies to increase cereal crop yield. Tree crops will require a more substantial increase in yield and reliability of yield as the Green Revolution gains in cereal crop productivity have not been implemented in tree crops. Adding to this challenge, the effects of climate change on yield are also uncertain, with changes in temperature, rainfall patterns and CO₂ levels likely to significantly influence crop productivity (Aydinalp and Cresser, 2008).

One innovative change that is being implemented in some tree crop industries, e.g. olives (Olea europaea) in Europe (Lo Bianco et al., 2021), is the transition to higher density orchard plantings. This transition requires the use of small trees with low shoot vigour and, ideally, precocious flowering (i.e., flowering at a younger age). This is often achieved through orchard management but can also be achieved using varieties and/or rootstocks that confer these traits. However, some tree crops like avocado (Persea americana) and macadamia (Macadamia integrifolia, Macadamia tetraphylla, and hybrids) lack available varieties or rootstocks with these traits, and many horticultural industries have not made substantive use of available varieties and/or rootstocks. Planting at higher densities gives the potential for higher yield per land as seen in olive orchards (Sobreiro et al., 2023), however, this may not be applicable to all tree crops (Haque and Sakimin, 2022). The benefits of higher density plantings are increased if trees flower and produce high yield at a younger age as this improves the return on investment and allows for faster turn-around of plantings and quicker introduction of newer, improved varieties. Another advantage of higher density plantings is that they are more amenable to implementation of automation/robotics for crop management and harvest. This is especially pertinent in countries with high wages and labour scarcity (Rutledge and Taylor, 2023) that often cannot compete on a cost level with countries with much lower wages. The use of automation can dramatically reduce the ongoing costs associated with crop management and harvest to make these industries remain competitive on an international standing.

Although responding to these challenges will require a multi-pronged approach, it is imperative that genomic technologies are harnessed to complement traditional breeding efforts to enhance and accelerate the process. The advent of low-cost genome sequencing and associated increase in high quality genome assemblies and ability to re-sequence large-scale germplasm and hybrid populations (Okemo et al., 2022) has dramatically advanced the implementation of genomic approaches into breeding programs. In particular, high-throughput sequencing enables the use of genome-wide association studies (GWAS) and marker-assisted selection (MAS) for high-throughput assessment of breeding populations to identify preferred parental crosses and filter progeny for traits of interest. However, as many tree crops are highly heterozygous outcrossers (Miller and Gross, 2011), traditional crossing and selection remains challenging as many desirable traits can be lost during crossing. The ability to genetically manipulate crops for trait improvement using newer gene editing technologies such as CRISPR holds the promise of genetically engineering traits without the need for genetically modified (GM) cultivars (Hu and Gao, 2023). These technologies allow for the genetic manipulation of individual genes while maintaining other desirable traits, although, as we will discuss later, many challenges remain. And while these approaches have been extensively used in many cereal crops, their use in most tree crops has been relatively unexplored and underutilised to date.

In this review we describe several key traits and genetic targets ( Figure 1 ) that are a focus for genetic improvement in tree crops in response to the challenges raised above. First, we explore the intensification of crop density through the modification of different aspects of tree architecture, including branching number, branching angle, and tree height. Next, we discuss bud dormancy in deciduous trees and how changes in temperature due to climate change may affect this important biological process. Following this we consider early, or precocious, flowering to accelerate tree crop breeding. And finally, we explore alternate bearing, a phenomenon experienced in many tree crops that influences yield predictability. While these traits are key targets for genetic improvement, there are many other important traits that are being targeted for improvement by breeding programs (see examples in Figure 2 ). In this review we also discuss existing and emerging genomic technologies that can be utilised for accelerating genetic improvement in tree crops. These rapid and precise genome-enabled technologies first require an understanding of the genetic basis of important traits in tree crops to inform marker assisted and genomic selection models or genetic manipulation of the tree crops through transgenics or gene editing.

Due to their extended lifespan, trees are in constant need of modifying their growth in response to the surrounding environment (Cooke et al., 2012). Tree architecture is an excellent example of this developmental plasticity, which is maintained via the regulation of axillary bud dormancy (Wang M. et al., 2019). Following establishment, axillary buds are maintained in a dormant state through a combination of various environmental and endogenous signals acting to ensure dormancy breaks at an appropriate time resulting in bud outgrowth and, if appropriate signals persist, in shoot branch formation (Wang M. et al., 2019). This high reliance of plant plasticity on environmental cues is increasingly becoming an issue in the face of rapidly changing, unreliable climatic conditions. Research focused on deciphering the genetic and molecular components/mechanisms regulating traits like shoot architecture and flowering is therefore critical to our ability to mitigate the impact of climate change on tree growth and productivity. Here, we will outline current knowledge of several key genetic components in tree crops regulating three specific shoot architecture traits: branching number, branching angle, and plant height. For horticultural crops, trees with shorter or less branches, and semi-dwarf phenotypes may be the ideal architecture allowing for higher planting density and reduced mechanical intervention like pruning (Scorza, 2005; Hollender and Dardick, 2015).

Shoot architecture has been a key target of domestication in many crops, with reduced branch number being an important component of the ideal shoot architecture ( Figure 1A ). In crops such as maize (Zea mays spp. mays), a decrease in tiller number correlates with yield increases (Doebley et al., 1997). The wild ancestor of maize (Zea mays spp. parviglumis) has an increased axillary branching phenotype (Doebley et al., 1995) and the gene controlling this trait was identified by Doebley et al. (1997) as TEOSINTE BRANCHED1 (TB1) which belongs to the TEOSINTE BRANCHED1, CYCLOIDEA, PROLIFERATING CELL FACTOR (TCP) family of transcription factors. Since then, orthologs have been identified and studied in fruit trees such as peach (Prunus persica) where overexpression of PpTCP18, a TB1/BRANCHED1 (BRC1) ortholog, in Arabidopsis thaliana caused a decrease in rosette branch numbers (Wang X. et al., 2023). Interestingly, in Carrizo citrange (Citrus sinensis x Poncirus trifoliata) the BRC1 ortholog THORN IDENTITY1 (TI1) acts partially redundantly with the BRC2 ortholog TI2 to control shoot branching alongside thorn identity (Zhang et al., 2020).

Shoot branch angle is another important aspect affecting overall shoot architecture which can lead to trees having an upright and erect architecture, termed pillar or broomy, or a spread apart architecture, termed weeping ( Figure 1B ). LAZY1 acts as a negative regulator of the polar auxin transport (PAT) to decrease the response to gravity, and loss-of-function mutations in rice promote a branching phenotype that is more spread apart (Li et al., 2007; Yoshihara and Iino, 2007). In apple (Malus domestica), the MdLAZY1A-W allele has a single nucleotide substitution that changes the amino acid from leucine to proline resulting in a dominant weeping phenotype (Dougherty et al., 2023). In contrast, an in-frame stop codon mutation in LAZY1 in silver birch (Betula pendula) leads to a recessive weeping phenotype (Salojärvi et al., 2017). LAZY1 belongs to the IGT gene family alongside another regulator of shoot branch angle, TILLER ANGLE CONTROL1 (TAC1). Mutations in rice (Oryza sativa) OsTAC1 result in a more compact phenotype (Yu et al., 2007; Jiang et al., 2012). The gene has since been identified in other species including the fruit trees peach (Dardick et al., 2013), plum (Prunus domestica; Hollender et al., 2018b), apple (Li et al., 2022), and sweet orange (Citrus sinensis; Dutt et al., 2022). In peach, the likely cause of the pillar phenotype in the ‘Italian pillar’ cultivar is an insertion causing a premature stop codon in PpeTAC1 (Dardick et al., 2013). Overexpression of PpeTAC1 in plum increased the branch angle, while RNA interference (RNAi) of PdoTAC1 in plum reduced the branch angle leading to a more upright architecture (Hollender et al., 2018b). Similarly, sweet orange CsTAC1 knockout transgenic lines displayed a more compact branch angle, and this was associated with low auxin levels and upregulation of CsBRC1 expression, following a similar expression pattern to Arabidopsis (Dutt et al., 2022). Another regulator of branching orientation identified in peach is the WEEP gene, encoding a sterile alpha motif (SAM) domain protein. The shoots on WEEP trees initially grow upwards but, after reaching a certain height (20 cm), start exhibiting a weeping phenotype (Hollender et al., 2018a). Grafting of buds from WEEP trees to wild-type rootstocks did not rescue the weeping phenotype of the scion suggesting that there is no requirement for a mobile or systemic signal for the weeping phenotype. Transcriptomic analysis showed differential expression of auxin-related genes like AUXIN/INDOLE-3-ACETIC ACID (Aux/IAAs) and AUXIN RESPONSE FACTORs (ARFs) indicating a role for auxin in the gravitropic response (Hollender et al., 2018a). RNAi of the WEEP gene in plum resulted in a weeping phenotype (Hollender et al., 2018a), showing conserved function of WEEP in a closely related tree species.

Plant height is another key trait that influences overall shoot architecture ( Figure 1C ). In the 1960s, a specific type of architecture with short internodes and fewer axillary branches was observed in the mutant apple McIntosh Wijcik (Fisher, 1969). Further research identified a dominant 2-oxoglutarate-dependent dioxygenase (DOX) gene, Columnar (Co) or MdDOX-Co, as the likely candidate responsible for the mutant phenotype (Lapins, 1976; Okada et al., 2016). Overexpression of MdDOX-Co in tobacco (Nicotiana tabacum) resulted in a dwarfed phenotype with a shorter main stem and internodes (Okada et al., 2020). This was associated with decreased endogenous bioactive gibberellin (GA) levels and exogenous application of GA was able to rescue the dwarfed phenotype (Okada et al., 2020). Overexpressing MdDOX-Co in Arabidopsis led to the discovery that it regulates 12-hydroxylation of GA₁₂, a precursor in the GA biosynthesis pathway, to reduce active GA levels resulting in the dwarf phenotype (Watanabe et al., 2021). Similar to the McIntosh Wijcik mutant, the peach dw mutants have a dwarfing phenotype associated with short internodes (Scorza, 1984). Mapping analysis identified a GA receptor gene GIBBERELLIN INSENSITIVE DWARF1c (GID1c) as the likely cause of the dw trait, and RNAi-induced silencing of PpeGID1c in plum led to varying degrees of dwarfing (Hollender et al., 2016). Rice breeding has for many years employed the use of the semidwarf (sd-1) allele, caused by a defective GA biosynthetic gene GIBBERELLIN 20-OXIDASE (GA20ox) to decrease overall size while increasing yield (Spielmeyer et al., 2002). Multiple GA biosynthesis genes were identified as differentially expressed between the ‘Williams’ banana (Musa acuminata) 8818-1 wild-type cultivar and its dwarf counterpart (Chen et al., 2016), and there are already cases of successfully using GA20ox gene(s) as a target to produce dwarf fruit varieties, such as via suppression of MpGA20ox1 in apple (Bulley et al., 2005) and via CRISPR-based silencing of five MaGA20ox2 genes in banana (Shao et al., 2020). Ethylene is another plant hormone that has been associated with plant height regulation. Overexpression of the lemon (Citrus limon) gene 1-AMINOCYCLOPROPANE-1-CARBOXYLIC ACID SYNTHASE 4 (CiACS4), involved in the ethylene biosynthesis pathway, in tobacco and lemon resulted in dwarf plants (Chu et al., 2023). CiACS4 was found to be directly upregulated by an ethylene response factor, CiERF023, and the CiACS4 protein also interacted with another ethylene response factor, CiERF3, to directly regulate the GA biosynthesis genes CiGa20ox1/2.

Here, we have discussed three main routes to shoot architecture manipulation: shoot branching numbers, branching angle, and/or overall plant height. The genetic regulation of these traits shows high levels of conservation between divergent plant species, thus providing several potential targets for improving architecture in other tree crops. Efforts to implement breeding and genomics-enabled approaches to alter these traits will provide significant inroads to achieving ideal tree architecture with maximised yield.

Deciduous trees that grow in temperate regions experience fluctuating temperature and daylength due to annual seasonal changes and have evolved to adapt to these environmental conditions by modulating their annual cycle of active growth and growth cessation. In deciduous trees, both terminal and axillary buds (whether vegetative or reproductive) temporarily stop visible growth during winter and only resume growth and flowering when the season is favourable (Singh et al., 2017). Growth cessation, bud set, cold acclimatisation and establishment of bud dormancy occur in sequential order with some overlap (Singh et al., 2017) creating difficulty not only in delineating where growth cessation ends and dormancy starts, but also in differentiating phenotypes and genetic mechanisms that are associated with these processes. In the state of deepest dormancy commonly referred to as endodormancy ( Figure 1D ; Lang et al., 1985), growth is repressed and cannot be initiated by growth-promoting conditions until the buds accumulate a certain number of hours of cold temperature known as the chilling requirement. Although this is a gradual process (Cooke et al., 2012), the transition to the ability of the plant to resume growth in permissive conditions is often referred to as ecodormancy (Lang et al., 1985). Release from dormancy is followed by budbreak, primarily driven by accumulation of heat (Alburquerque et al., 2008). Thus, dormancy is an important physiological and adaptive process that helps plant buds survive harsh winter temperatures and is an important determinant of yield in deciduous tree crops.

Chilling and heat requirements are specific to the plant species and genotype, and reflect the long-term adaptation to specific growing conditions in their native or cultivated regions (Luedeling, 2012). For example, the specific phenology of the almond (Prunus dulcis) cultivar ‘Nonpareil’ grown in South Australia is depicted in Figure 3 . Delayed budbreak and low yields arising as a consequence of reduced winter chilling in the warming climate are posing a global threat to agriculture and food security (Luedeling, 2012; Atkinson et al., 2013). For this reason, tree phenology has been a subject of extensive research, with significant progress made in understanding the environmental, physiological, genetic, epigenetic and hormonal regulation of dormancy (Lloret et al., 2017; Beauvieux et al., 2018; Liu and Sherif, 2019; Fadón et al., 2020; Pan et al., 2021; Yamane et al., 2021; Yang Q. et al., 2021; Nilsson, 2022). Not only do we need in-depth understanding of the molecular mechanisms regulating dormancy, but also taking advantage of next generation molecular techniques and translating these advancements to the orchard is crucial for plant biologists to develop new and improved cultivars better suited to the new climate dynamics.

In Rosaceous plants like almond, peach, plum, pear (Pyrus spp.) and apple, DORMANCY ASSOCIATED MADS-BOX (DAMs) transcription factors are central regulators of bud dormancy during winter chilling ( Figure 1D ). DAMs were first identified in a natural mutant peach cultivar evergrowing, which fails to enter dormancy during winter (Bielenberg et al., 2008). Sequencing revealed six genes that were either partially or fully deleted in the mutant compared to wild-type peach and were termed as DAMs (Bielenberg et al., 2008). Various studies in apple (Wu et al., 2017), Japanese pear (Pyrus pyrifolia; Saito et al., 2013; Niu et al., 2016; Tuan et al., 2017) and Japanese apricot (Prunus mume; Yamane et al., 2019) have shown that overexpressing DAMs prolongs dormancy in these trees, and silencing or mutagenesis prevents dormancy in hybrid aspen (Populus tremula x P. tremuloides; Singh et al., 2019) and apple (Moser et al., 2020; Wu et al., 2021). A recent study by Falavigna et al. (2021) explored the gene regulatory networks controlled by DAMs during the dormancy cycle in apple in response to environmental cues and hormonal signalling pathways. Transcriptomic studies in various other species including WT-717 poplar hybrid, Japanese pear and raspberry (Rubus idaeus) show that DAMs, as well as important genes involved in abscisic acid (ABA) and GA synthesis and catabolism differentially accumulate in buds transitioning from growth cessation through the different stages of dormancy to growth resumption in spring (Mazzitelli et al., 2007; Ruttink et al., 2007; Niu et al., 2016). ABA has been shown to control dormancy by inhibiting genes involved in cell proliferation and growth, and activating catabolism genes of other hormones such as GAs (Yang Q. et al., 2023). However, the exact signalling cascades by which DAMs trigger ABA and GA pathways in response to environment are still obscure.

While DAMs play an important, conserved role in regulating bud dormancy in many temperate crops, other genes are also likely to play an important role in this process. For example, in apple the major quantitative trait locus (QTL) controlling bud dormancy does not contain any DAM genes suggesting other gene(s) are regulating bud dormancy in apple (Allard et al., 2016). The C-REPEAT BINDING FACTOR1 (PpCBF1) gene has been shown to play a role in bud dormancy when overexpressed in apple (Wisniewski et al., 2015). While CBF and other genes likely play a role in bud dormancy, the DAMs are clearly critical regulators of the onset and release of bud dormancy in Rosaceae species. Delving deeper into the downstream targets regulated by DAMs can reveal new targets towards modifying bud dormancy according to the geographical climates of orchard production.

A critical developmental time in the life cycle of a tree is the vegetative to reproductive phase change, during which the tree attains the ability to flower and produce seeds. When compared to an annual plant, which can flower, produce seeds, and complete its life cycle within one year, a woody perennial tree will often not reach reproductive maturity for several years or even decades (Hackett, 1985). Following the first onset of flowering, trees flower annually in response to environmental signals but commit only some of the meristems to flowering, thereby maintaining the polycarpic growth habit (Amasino, 2009). The delayed maturity, long life span, and polycarpic growth of trees require a complex regulatory network to regulate the timing of developmental transitions and synchronise environmental signals with phenological events to ensure survival and successful reproduction (Brunner et al., 2017). While flowering gene homologs from woody perennials often show a conserved role in regulation of flowering in model annuals such as Arabidopsis and tobacco, functional studies in trees reveal shared and distinct functional aspects, highlighting the diverse roles of paralogs (Hsu et al., 2011; Sheng et al., 2022) and the pleiotropic effects of flowering genes on growth (André et al., 2022; Sheng et al., 2023), organ identity (Zhang et al., 2021), and phenology in trees (Ding and Nilsson, 2016). Therefore, targeting specific genes to reduce the years to reproduction in trees is a complex task that depends on the species in question. However, the types of genes that could be targeted to accelerate reproductive maturation are those regulating flowering time and floral meristem identity ( Figure 1E ).

Members of the phosphatidylethanolamine–binding protein (PEBP) family are crucial for floral transition. The divergence of PEBP genes into FLOWERING LOCUS T (FT) (predominantly activators of flowering) and TERMINAL FLOWER1 (TFL1)/CENTRORADIALIS (CEN) (predominantly repressors of flowering) (Wickland and Hanzawa, 2015) subgroups in angiosperms was critical for the reproductive success of flowering plants (Shalit et al., 2009; Karlgren et al., 2011). Further divergence through gene duplication events, followed by neo- and sub-functionalisation, gave rise to FT-like regulators implicated to have broader functions in plant development and adaptation (Pin and Nilsson, 2012). It is important to note that the specific functions and regulatory mechanisms of FT in perennial fruit trees can vary between different species and cultivars within the same species, but FT has generally maintained its role as an inducer of flowering. Early flowering can be induced in Arabidopsis by overexpressing FT orthologs from many diverse tree crops such as avocado (Ziv et al., 2014), mango (Mangifera indica; Fan et al., 2020), apple (Kotoda et al., 2010), kiwifruit (Actinidia chinensis; Voogd et al., 2017), and litchi (Litchi chinensis; Ding et al., 2015); while constitutive ectopic expression of FT promotes early flowering in tree crops such as apple (Kotoda et al., 2010), trifoliate orange (Poncirus trifoliata; Endo et al., 2005), and kiwifruit (Voogd et al., 2017). However, in some instances, overexpression of FT causes in vitro and abnormal flowering such as that observed in apple (Kotoda et al., 2010) and kiwifruit (Voogd et al., 2017; Moss et al., 2018). More recently, it was shown that FT chimeric proteins induce floral precocity without detrimental effects in citrus (Sinn et al., 2021) and kiwifruit (Herath et al., 2023) species, suggesting that this approach may be universally applicable for woody perennial crops.

The concept of grafting on rootstocks engineered to overproduce FT to accelerate flowering of the scion has been explored as means of flowering control in horticultural and agricultural practices. However, the movement of FT and its ability to promote flowering in trees is still under debate (Putterill and Varkonyi-Gasic, 2016), with reports of movement across the graft union without precocity observed in juvenile apple, trifoliate orange, hybrid aspen or WT-717 poplar hybrid scions (Zhang et al., 2010; Wenzel et al., 2013; Freiman et al., 2015; Miskolczi et al., 2019; Wu et al., 2022), while precocious flowering was observed in Carrizo citrange (Soares et al., 2020) and Jatropha curcas scions (Ye et al., 2014). In various Jatropha spp., the distance between the graft union and bud is important for floral induction and increasing this distance can lead to a lower frequency of flowering (Tang et al., 2022). Therefore, grafting approaches may be dependent on many variables including the plant species, plant/scion size, as well as the FT gene and its regulatory promoter.

Plant viral vectors have also been used to either express FT for virus-induced flowering or silence TFL1 via virus-induced gene silencing (VIGS), and these approaches have induced early flowering in apple, Japanese pear, European pear (Pyrus communis), and Vitis spp. (Sasaki et al., 2011; Yamagishi et al., 2014; Velázquez et al., 2016; Yamagishi et al., 2016; Maeda et al., 2020). RNA silencing of TFL1/CEN has also induced early flowering in apple (Kotoda et al., 2006), WT-717 poplar hybrid (Mohamed et al., 2010) and European pear (Freiman et al., 2012). Recently, CRISPR/Cas9-mediated mutagenesis of one or two CEN genes induced fast flowering in several kiwifruit species (Varkonyi‐Gasic et al., 2019; Varkonyi-Gasic et al., 2021; Herath et al., 2023) and editing of TFL1 induced early flowering in apple and European pear (Charrier et al., 2019). However, editing of a blueberry (Vaccinium corymbosum) CEN gene did not result in precocious flowering (Omori et al., 2021), suggesting that some TFL1/CEN genes perform different roles. Indeed, mutagenesis of a kiwifruit BROTHER OF FT AND TFL1 (BFT) prevented the establishment of dormancy without affecting flowering (Herath et al., 2022), and in Carrizo citrange loss of CsCEN resulted in the conversion of axillary meristems to thorns, while ectopic CsCEN expression converted thorns to axillary meristems (Zhang et al., 2021).

Floral meristem identity genes perform as key outputs of FT-mediated floral transition or FT-independent flowering pathways. They include MADS-box proteins such as APETALA1 (AP1) and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 (SOC1), necessary to initiate and maintain reproductive development, and LEAFY (LFY) acting as a pioneer transcription factor reprogramming vegetative meristematic cells into an inflorescence (Jin et al., 2021; Lai et al., 2021). LFY homologs from trees, including peach, litchi and mango, can induce flowering when overexpressed in annual models (An et al., 2012; Ding et al., 2018; Wang Y. et al., 2022). Ectopic overexpression of LFY also promoted flowering in Carrizo citrange (Peña et al., 2001), yet there were some inconsistencies in floral induction between various Populus hybrids that displayed altered flower development as well as sexual differentiation (Weigel and Nilsson, 1995; Rottmann et al., 2000). MADS-box genes have been important targets for selection during crop domestication and improvement as they play pivotal roles in every aspect of plant reproductive development, controlling flowering time, inflorescence architecture, determination of floral meristem and floral organ identity, and seed development (Schilling et al., 2018). The ability of AP1 homologs to induce flowering in a range of woody perennials, e.g. Carrizo citrange, silver birch and apple (Peña et al., 2001; Elo et al., 2007; Flachowsky et al., 2007), has been adopted for rapid cycle breeding approaches in apple and European pear (Flachowsky et al., 2011; Tomes et al., 2023). However, functional studies in hybrid aspen revealed that, like FT, homologs of MADS-box genes controlling the regulation of flowering time in annual plants have a role in control of vegetative phenology in trees (Azeez et al., 2014; Ding and Nilsson, 2016).

This section has highlighted several key genes and gene families that offer valuable tools for shaping and improving developmental and reproductive traits in crops. Their manipulation through targeted breeding approaches can contribute to the accelerated development of new crop varieties with enhanced characteristics (as demonstrated in Figure 4 ), ultimately benefiting horticultural food production.

Alternate bearing, also known as biennial bearing, is a significant issue faced by the horticultural industry and has been reported to affect mango, cranberry (Vaccinium macrocarpon), avocado, olive, apple, European pear, plum, apricot (Prunus armeniaca), coffee (Coffea arabica), Citrus spp., litchi and nut producing trees (Monselise and Goldschmidt, 1982). Alternate bearing is initiated by an abnormally heavy fruit load one year, decreasing the ability of the tree to undergo floral initiation and resulting in a light fruit load the following season. This pattern of alternating high then low crop production years can differ in intensity between cultivars within a species, and can also be strongly influenced by environmental factors (Monselise and Goldschmidt, 1982). These inconsistent patterns of bearing and yield have major implications for growers and for responding to the need for reliable food sources into the future.

Studies have demonstrated that fruit load affects flowering the following season by repressing the expression of genes involved in floral initiation ( Figure 1F ). FT, AP1 and LFY are shown to be highly expressed in the absence of fruit and have minimal to no expression in the presence of a heavy crop load in avocado (Ziv et al., 2014), mango (Nakagawa et al., 2012; Das et al., 2019), and ‘Moncada’ mandarin (Citrus clementina x Kara mandarin (C. unshiu x C. nobilis); Muñoz-Fambuena et al., 2011). In avocado, expression of floral initiation genes was suppressed in shoots with high fruit load, and the transition from vegetative to reproductive structure was completely repressed with young leaves appearing instead (Ziv et al., 2014). In Citrus spp., CiFT2 expression is downregulated in trees with high fruit load by CcMADS19, an ortholog of the floral repressor FLOWERING LOCUS C (FLC), while in trees with a low fruit load, low CcMADS19 expression results in an increase in CiFT2 expression (Agustí et al., 2020). Expression of CcMADS19 itself was shown to be correlated with epigenetic changes (Agustí et al., 2020). Fluctuations in activity of SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL) genes have also been observed in response to fruit load in mandarin (Citrus reticulata; Shalom et al., 2012), mango (Sharma et al., 2020), and apple (Guitton et al., 2016).

Studies in ‘Murcott’ mandarin (Citrus reticulata x C. sinensis) and olive have shown the possible involvement of auxin (IAA) in alternate bearing, with auxin transmitting a heavy-fruit-load signal to the shoot apical meristem to repress floral initiation, and thus playing a role in determining whether vegetative or floral development proceeds (Haim et al., 2021). Peaks in transcript levels of IAA transporter genes were also associated with the induction of the flowering-inhibition signal (Haim et al., 2021). It has been shown that auxin induces the biosynthesis of GA during fruit set in tomato (Solanum lycopersicum; de Jong et al., 2009; Hu et al., 2018) and woodland strawberry (Fragaria vesca; Liao et al., 2018), and GA has been widely shown to promote vegetative growth and inhibit flower formation in tree crops. Heavy fruit load increases GA levels in ‘Satsuma’ mandarin (Citrus unshiu) leaves (Koshita et al., 1999), and fruit load affects the expression patterns of GA metabolism genes in mango (Nakagawa et al., 2012; Sharma et al., 2020). Exogenous application of GA also leads to a significant increase in TFL1 expression in apple (Haberman et al., 2016) and reduced FT expression in ‘Orri’ mandarin (Citrus reticulata x C. temple; Goldberg-Moeller et al., 2013), apple (Elsysy and Hirst, 2019), and mango (Krishna et al., 2017). Therefore, in fruit trees, auxin likely plays a role in mediating fruit-load inhibition of flowering by increasing GA levels to regulate expression of both floral promoter and repressor genes.

Another important factor regulating alternate bearing is resource availability, in particular, carbohydrate levels. Competition for carbohydrates between developing fruit and nearby apical buds leads to depletion of carbon levels and reduced cellular activity in vegetative meristems, blocking the onset of floral development and potentially inducing bud dormancy (Tarancón et al., 2017; Martín-Fontecha et al., 2018). Tree starch content is significantly depleted by a heavy crop load resulting in fewer flowers and a lower crop load the following year (Nakagawa et al., 2012). Sugars are increasingly recognised as important signalling molecules (Fichtner and Lunn, 2021) with several sugar signalling pathways associated with bud outgrowth. In mandarin, buds on branches with high fruit-load showed induction of the trehalose metabolism enzymes, TREHALOSE PHOSPHATE PHOSPHATASE (TPP) and TREHALOSE PHOSPHATE SYNTHASE (TPS), which play a role in flowering in model species (van Dijken et al., 2004), suggesting a possible role for trehalose and/or its biosynthetic pathway in the transition to flowering (Shalom et al., 2012). The promotion or repression of flowering may be regulated by changes in sugar signalling which reflects changes in resource availability i.e., when many fruit are present, buds are deprived of photoassimilates, preventing flower bud formation.

By improving our understanding of the molecular regulation of alternate bearing we can continue to develop and optimise targeted tree management strategies, such as plant growth regulator application, nutritional management and regulating carbohydrate depletion, to ensure the consistent promotion of flowering and fruit development. Additionally, the significant variation in susceptibility to alternate bearing observed between cultivars offers potential targets for genetic manipulation to reduce alternate bearing and the need for often time-consuming tree management strategies.

Plant breeding has been transformed in recent years through the advent of rapid and precise genome-enabled breeding technologies. These cutting-edge methods have enabled breeding programs to accelerate genetic gain using techniques that accelerate the development of superior crop varieties ( Table 1 ). In the following section we examine the spectrum of genome-enabled plant breeding methods. First, we delve into purely genomic techniques, such as MAS and genomic selection (GS). Both methods use the power of molecular markers and genomic data to identify and select plants with desirable traits more efficiently. By doing so, they can dramatically shorten breeding cycles and reduce the size of resource-intensive progeny trials. Moving beyond MAS and GS, we explore how advances in functional genomics and computational biology enable plant scientists to unravel the complex web of genetic interactions governing plant traits. Greater understanding of gene regulatory networks provides precise information into which genes are best to target for genetic modification, allowing molecular breeders to fine-tune desirable traits with precision. Finally, we review genetic manipulation in plants, evaluating the potential for transgenic and gene editing approaches to precisely modify traits in elite crop varieties. Most of these technologies are reliant on, or aided by, a sound understanding of the genetic basis of traits of interest, such as described in the previous sections.

The rich history of traditional horticultural tree breeding has domesticated a diverse variety of fruits, refined through generations of crossing and selection. This process has generated high-yielding citrus, apple, almond and avocado cultivars, among others, that have become the staple fruits produced in orchards around the world. Yet, in the face of evolving challenges and opportunities posed by shifting climatic patterns, the growing demand for food and the emergence of automation, the limitations of conventional breeding methods are becoming increasingly apparent. To effectively address these threats to horticultural production and to rapidly develop resilient varieties, there is a need for greater integration of genome enabled technologies into horticultural tree breeding programs. Apple breeding is one of several industries that has led the way for integration, showcasing remarkable success in expediting the development of novel cultivars boasting enhanced disease resistance, drought resilience, and superior fruit quality. The advent of low-cost, large-scale genome sequencing and the expanding repertoire of genomic approaches, such as those outlined in this review, are shifting the research and development landscape in tree crop species enabling greater integration of genome-enabled technologies across the horticultural tree crop breeding sector. By embracing genome enabled technologies in traditional breeding programs, we not only increase our capacity to develop new varieties, but also produce tree crops better suited to existing and emerging challenges, safeguarding the future of next generation orchards.

SK: Conceptualization, Writing – original draft, Writing – review & editing, Visualization. SS: Writing – original draft, Writing – review & editing. LP: Writing – original draft, Writing – review & editing, Visualization. MM: Writing – original draft, Writing – review & editing. LS: Writing – original draft, Writing – review & editing. AJ: Writing – original draft. JV: Writing – original draft, Writing – review & editing. MT: Writing – review & editing. CC: Conceptualization, Writing – review & editing, Writing – original draft. EV-G: Conceptualization, Writing – original draft, Writing – review & editing, Visualization. PP: Writing – original draft, Writing – review & editing.