Advances in sequencing and key character analysis of mango (Mangifera indica L.)

Abstract Mango (Mangifera indica L.) is an important fruit crop in tropical and subtropical countries associated with many agronomic and horticultural problems, such as susceptibility to pathogens, including powdery mildew and anthracnose, poor yield and quality, and short shelf life. Conventional breeding techniques exhibit significant limitations in improving mango quality due to the characteristics of long ripening, self-incompatibility, and high genetic heterozygosity. In recent years, much emphasis has been placed on identification of key genes controlling a certain trait through genomic association analysis and directly breeding new varieties through transgene or genotype selection of offspring. This paper reviews the latest research progress on the genome and transcriptome sequencing of mango fruit. The rapid development of genome sequencing and bioinformatics provides effective strategies for identifying, labeling, cloning, and manipulating many genes related to economically important traits. Preliminary verification of the functions of mango genes has been conducted, including genes related to flowering regulation, fruit development, and polyphenol biosynthesis. Importantly, modern biotechnology can refine existing mango varieties to meet the market demand with high economic benefits.


Introduction
Mango (Mangifera indica L.) is a juicy drupe of Mangifera of Anacardiaceae. It is the world's third most planted tropical fruit after banana and pineapple (http://www.fao.org/faostat/), widely grown in tropical and subtropical marginal areas [1]. The main planting areas are hillsides, river valleys, or wilderness forests at an altitude of 200-1350 meters in India, China, Thailand, Myanmar, Bangladesh, and Malaysia. The popularity of mango can be attributed to its attractive taste, fragrance and high nutritional value. The fruit consists of pulp, peel, and kernel. The pulp is rich in reducing sugars, amino acids, aromatic compounds, and functional compounds such as: pectin vitamins, anthocyanins, and polyphenols [2]. The β-carotene in mango f lesh is as high as 200 mg/100 g, which is 10 and 50-fold higher than in bananas and apples. Mangiferin is the main active component in mango leaves and helps to scavenge oxidative free radicals and participate in antibacterial and immunomodulation [3,4].
There are more than 1000 mango cultivars worldwide [5,6]. According to its embryo type, mango fruit can be divided into monoembryonic and polyembryonic varieties. The seeds of the monoembryonic (India) variety have only one zygotic embryo, which is propagated by sexual reproduction, and only one seedling after sowing. The seedlings exhibit great variability but do not maintain the excellent characteristics of the female parent, mainly distributed in subtropics, and the peel is mostly red. In contrast, the polyembryonic (Southeast Asia) type is most common in the tropics, and the polyembryonic traits are often dominant. The polyembryonic types are produced from mother plants, and several seedlings can grow after sowing. Embryos that can develop into seedlings are mostly asexual, accounting for the small variability of fruiting trees, and most can preserve the traits of mother plants. The pericarp is mainly green to yellow, and Thai mango (Mangifera siamensis warbg. ex Craib) mostly belongs to this type [7][8][9]. With the development of mango variety breeding, the hybridization between monoembryonic and polyembryonic types can yield polyembryonic offspring.
Evaluating and protecting natural mango germplasm resources is essential, and new varieties are warranted for the modern market and commercial needs. Most mango plants are heteroecious and cross-pollinated. According to literature records, the Indian Agricultural Research Association (IARI) first conducted breeding research to improve mango varieties in 1961. At present, seed selection, cross-breeding, and mutation breeding remain the main breeding ways. Due to the long juvenile phase of perennial fruit trees, the selection of mango breeding offspring by the above methods is time-consuming and requires a lot of screening work, and the characteristics of hybrid offspring are often significantly different. Plant biotechnology provides a new approach for improving varieties and developing stable and efficient genetic transformation methods. It has become an important means for genetic improvement and germplasm resource innovation of mango fruit.
Over the years, the mango hybrid offspring of different parent varieties have been used to construct genetic populations.
Twenty-seven hybrids of Tommy Atkins, Haden, Palmer, Coquinho, Kent, and Van Dyke varieties have been used to estimate quality genetic parameters by modeling [10] and evaluating mango hybrids obtained through open pollination based on physical and chemical traits of the fruit [11]. Hybrid populations can help in genetic diversity analysis with significant emphasis placed on quantitative trait locus and marker-assisted selection. The offspring of JinHwang and Irwin was used to construct the first high-density genetic map for high-throughput sequencing and specific-locus amplified fragment (SLAF) library construction [12]. Interestingly, mining of RAPD primers in Indian mango hybrids has recently been conducted [13]. Besides, new hyper-variable mango SSRs (MSSRs) designed from Amrapali genome sequences have been used to discover polymorphisms between Amrapali and Sensation parental genotypes [14], and single nucleotide polymorphism (SNP) markers to genotype mango hybrid populations [15].
Early trait prediction is essential to shorten the breeding process. Molecular markers are genetic markers based on the polymorphism of biological macromolecules, especially the genetic material (nucleic acid) of organisms. They are directly expressed in the form of DNA and are not interfered with by tissue types, development periods, environmental conditions, etc. Compared with the morphological traits studied by traditional genetics and breeding, they exhibit obvious advantages, mainly ref lected in the fact that the stage of plant development, gene expression and environmental changes do not affect the selection of target traits, easier and faster to overcome undesirable trait linkage and introduce distant superior genes [16]. Over the years, simple sequence repetitions (SSRs) markers [17], cleavage amplification polymorphic sequence markers [18], and reverse transposon insertion polymorphic markers [19] have been widely used in mango research. The past decade has witnessed significant progress and heralded the era of genome breeding. The key genes controlling a certain trait can be located by high-throughput sequencing and association analysis, and new varieties are bred directly by transgene or genotype selection of offspring. Accordingly, the de novo transcriptome data [20][21][22] and the genetic map data [12,23] of mango have been successively published. In 2020, Wang et al. conducted high-depth whole genome sequencing of mango and documented the whole genome data at the chromosome level [22]. Modern biotechnology is an effective adjunct to traditional mango breeding. This paper will focus on the present situation and prospect of genetic improvement of mango using molecular and biotechnology. Importantly, more advanced biotechnology tools and synthetic biology will provide efficient gene editing means for improving agronomic mango traits in the future.

Fine-scale genomic map of mango
Before second generation sequencing technology was first used to sequence mango in 2014, mango genome sequence resources were very scarce, with only 684 highly redundant sequence entries in the GenBank. Azim et al. first sequenced, assembled, and annotated the mango chloroplast genome. The chloroplast genome of mango was 151, 173 bp in size, comprising a pair of reverse repeats of 27, 093 bp, separated by large and small single copies. A total of 91 out of 139 genes in the mango chloroplast genome were protein-coding genes. Sequence analysis showed that the chloroplast genome of citrus was closest to that of mango [24]. To better understand the basic molecular biology of mango fruit, large-scale discovery and characterization studies of functional genes by genome sequencing or transcriptome have been carried out worldwide. In 2018, Qamar-ul-Islam et al. reported four mango varieties -Cv. Langra, Cv. Zill, Cv. Shelly, and Cv. Kent -and, moreover, conducted a comparative analysis of the transcriptome [25]. This paper reports the world's first online genome resource focusing on mango. It contains predicted gene information of the whole genome, unigenes annotated by homologous genes of other species, and Gene Ontology (GO) terms, providing a mango genome resource and allowing users to analyse the genome database of four mango varieties for genetic improvement and management of mango genome [25].
Although mangoes are highly heterozygous, the current evidence suggests that the mango germplasm is diploid (2n = 2x = 40 chromosomes) [7,26]. In this respect, in 1991, Arumuganathan et al. analysed the nucleus DNA content of mango in their study using f low cytometry with propidium iodide staining of isolated nuclei, it was proposed that the genome size of mango is about 4.39 × 10 8 bp [27]. Besides, the number of chromosomes in mango somatic cells was determined by the Carbol fuchsin method to be 2n = 40, and no individuals with other ploidy were found [28,29]. The polyembryonic mango 'Gomera-1' has been confirmed to be diploid by f low cytometry and chromosome count analysis [30]. Yonemori et al. conducted the first study using f luorescence in situ hybridization (FISH) technique with 5S and 45S ribosomal DNA (rDNA) as probes on the mid chromosomes of mango somatic cells and discriminated 8 out of 40 chromosomes [31]. This information provides a basis for understanding the number of chromosomes mounted during sequencing.
In 2020, the first mango genome was published, providing a fine-scale mango genome map [22] with a size of 393 Mb by deep sequencing and assembly of data on the traditional mango variety Alphonso. From 2020 to 2022, genome and transcriptome analyses have been used to analyse variations in gene sequence and gene expression and specifically applied to fruit development-related research in mango ( Fig. 1, Table 1). At the same time, Li et al. obtained a 371.6 Mb genome from Hong Xiang Ya mango by SMRT sequencing, which contained 34 529 predictive protein-coding genes, providing the genetic basis for understanding special phytochemical compounds related to fruit quality [32]. Ma et al. sequenced Irwin varieties and obtained a high-quality genome sequence of 396 Mb. After transcriptome analysis, they found that the transcriptional regulation of the MiPSY1 gene was related to β-carotene biosynthesis during mango fruit ripening, which provided a genome platform for studying the molecular basis of mango f lesh color [33]. The Mango Genome Consortium (https:// mangobase.org/easy_gd b/index.php) sequenced, recombined, analysed, and annotated the genome of the monoembryonic mango variety Tommy Atkins and used the hybrid between Tommy Atkins and Kensington Pride to generate phased haplotype chromosomes and a high-resolution phased single nucleotide polymorphism map, beneficial to identify quantitative trait loci (QTL), gene and haplotype related to fruit weight [34]. In 2022, Cortaga et al. determined the whole genome sequences of three Philippine mango species (Carabao, Huani, and Paho) for identifying genome-wide specific markers for these Philippine native mango varieties [35]. This genomic information can guide future research to better understand the growth, survival status, and gene regulation mechanisms of mango.

Transcriptome sequencing analysis of mango
As shown in Fig. 1 and Table 2, current mango transcriptome research has mainly focused on the process of fruit development. Fruit ripening is a complex process during which the development of f lavor and color, the change and softening of cell wall components, the degradation of starch, and the development of aroma occur and determine the unique fruit characteristics. In 2014, Azim et al. used the short-read assembly program Trinity for de novo transcriptome assembly of Langra mango, which was the first report on the transcriptome of Rhizaceae plant members. More than 13 500 unigenes were assigned to 293 KEGG pathways. In addition to the main pathways related to plant biology, KEGG pathway analysis revealed significant enrichment in a series of biochemical pathways involving (i) biosynthesis of bioactive f lavonoids, f lavonoids and f lavonols; (i) biosynthesis of terpenoids and lignin; and (iii) plant hormone signal transduction, providing novel insights into exploring key regulatory genes in mango growth and development by transcriptome technology [24]. Fruit-specific secondary metabolites and aroma volatiles are important markers to distinguish between immature and mature stages, such as carotenoids, anthocyanins, and aroma in mango fruits. Transcriptome analysis of Alphonso mango fruit was used to analyse the unique transcription profile characteristics affecting fruit f lavor, color, ripening time, ripening pattern from peel to the core and long shelf life [48]. The transcriptome analysis of Amrapali mango was used to clarify the transcription trend of key genes related to peel color in the anthocyanin biosynthesis pathway. Among the 108 transcription sequences of the phenylpropanoid f lavonoids pathway, 15 contigs were identified as anthocyanin biosynthesis genes [43]. To explore the molecular basis of mango f lavor formation, the molecular determinants of carotenoid and aroma composition in mango [20,47] were explored using volatile spectrum, metabonomics, and transcriptomics. The MYB, bHLH, and NAC transcription levels were highly correlated with pulp pigment content, which may be related to carotenoid accumulation. This finding highlights the main differences in metabolic pathways during fruit ripening, which may lead to a change in mango fruit f lavor, and reveals several related genes for future studies. The discovery of transposonmediated ncRNA in crops has facilitated analysis of metabolic regulation in mango fruit, with around 100 miRNA and more than 7000 temperature-responsive lncRNA. Interestingly, some lncRNA-targeted miRNA could reduce the stability of lncRNA, and the target genes of these ncRNA were characterized. The newly identified mango ncRNAs may play potential roles in biological and metabolic pathways such as growth and development, pathogen defense mechanism, and stress response process [49].
Sweetness is an important trait that determines fruit quality. It is well-established that during mango f lesh ripening, starch is hydrolyzed into sucrose, fructose, and glucose with different concentrations catalyzed by invertase and β-glucosidase, which account for the unique sweet taste of mango varieties. Transcriptome analysis of differences in sugar accumulation between the high-sweet mango Tainong-1 and low-sweet Mango Renong-1 found that the key genes exerted a synergistic effect in sucrose transport, metabolism, and biosynthesis through regulating transcription factors such as MYB and NAC was the main reason for high sugar content, but no specific regulatory gene was identified [32].
Abiotic and biotic stresses are important factors affecting fruit ripening. Environmental stress conditions such as drought, salinity, high temperature, and f lood can significantly interfere with the development and yield of tropical fruit trees and affect the fruit quality. Ripening can change the hardness of fruits, making them vulnerable to pests and pathogens in the final stages of ripening or during storage [50]. Gene expression analysis was used to clarify the biological mechanism of hot water brushing (HWB) activation in mango regulating fruit quality [42] and resistance to postharvest diseases [37]. The results showed that a high temperature could induce internal tissue decomposition of mango fruit and synthesis of reactive oxygen species (ROS) at 44 • C and increase the expression of abiotic and senescencerelated genes in mango fruit in response to heat stress [37,45]. In a study on disease resistance and defense genes, Hong et al. obtained the first reference transcriptome data of Colletotrichum gloeosporioides in postharvest mango by high-throughput nextgeneration sequencing technology [39]. In the same year, Sela et al. discovered a new M. indica latent virus (MILV) virus sequence for whole transcriptome sequencing of mango fruit. Although no virus-related symptoms were detected, the differential gene analysis of mango peel transcriptome showed significant stress in mango peel, and the gene expression related to plant immune response to pathogen and virus infection increased [51].  [16]. The main advantage of developing molecular markers based on the transcriptome sequences is to increase the possibility of finding associations between functional genes and phenotypes and analysis of key traits, including fruit size, fruit f lavor and storability of hybrid offspring [16]. The first high-density genetic map of Mango was constructed by highthroughput sequencing of 173 F1 lines hybridized by JinHwang and Irwin: 6594 SLAFs were organized into a linkage map consisting of 20 linkage groups and were conducive to future genome assembly [12]. The discovery of RAD-based markers improves the development of network genome resources for plant genetic improvement and germplasm management and identifies SNP in the whole genome [53,54] [52,57], making genetic maps to identify genomic markers and regions related to important horticultural traits of mango (such as embryo type, branching habit, f lowering, peel/pulp color, and beak shape) [23], and further improving breeding efficiency. Genome-assisted breeding provides the necessary resources for developing high-density and cost-effective genotyping research, which is of great help to mango breeding and genome-wide association of yield and quality traits.

Functional verification of genes related to flowering regulation in mango
Using omics data, researchers have focussed on studying functions at the molecular level, identifying genes through annotation, studying expression regulation mechanisms and functions in metabolic pathways of organisms, analysing the relationship between genes and products, and predicting and  discovering protein functions. Current transcriptomics research of mango fruit has mainly focused on the fruit, while the functional verification of genes at physiological and biological levels has focused on f lowering regulation (Table 4), fruit development, and metabolite synthesis (Table 5). Flowering and fruit parameters play a critical role in growth and development events, and many genes and proteins related to f lowering regulation have been isolated and identified in mango (Fig. 2). The GIGANTE (GI)-(FKF1)-(CDF1)-(CONSTANS) CO module has been associated with the regulation of f lowering in the photoperiod and circadian pathways [92]. The expression of MiGI in mango is controlled by photoperiod and biological clock and forms a complex with the MiFKF1 protein to induce MiCO expression. The MiGI-MiFKF1 complex degrades MiCDFs, which inhibits the transcription of MiCO and MiFT genes [66]. Thirty-six CO homologous genes in mango were found by transcriptome data analysis [93]. Overexpression of MiCO, MiCOL1A, MiCOL1B, MiCOL16A, and MiCOL16B significantly delayed the f lowering time of transgenic Arabidopsis thaliana and enhanced the drought tolerance of transgenic A. thaliana, which may be due to inhibition of AtFT and AtSOC1 expression [61,63,69]. SVP gene is involved in mediating the environmental temperature, autonomic regulation and the vernalization pathway to regulate f lowering [94]. Moreover, mango MiSVP1 overexpression in A. thaliana can delay f lowering time by promoting AtFLC expression and inhibiting AtFT and AtSOC1 levels, while MiSVP2 overexpression can promote the expression of AtFT and AtSOC1, inhibit AtFLC expression and accelerate f lowering [68]. Three MiFTs homologous genes in mango were only increased in leaves under optimal f lower induction conditions, and treatment with 250-ppm gibberellin 3 (GA3) completely inhibited f lowering and MiFT expression under heavy crop load and no crop load conditions [58]. Tissue-specific expression patterns showed that MiFT1 expression increased sharply in leaves and was significantly higher than the other two MiFTs during f lower bud development. Overexpression of three MiFTs in A. thaliana showed that MiFT1 yielded the most potent effect on promoting f lowering [62].
MiTFL1 is involved in the regulation of f lowering mediated by the aging pathway, and overexpression of MiTFL1-1 and MiTFL1-3 in A. thaliana can affect the development of f lower organs [95]. There are four APETALA1 (AP1) homologous genes in mango, 2020) [65] No. GQ152892

MiGI2
Temperaturedependent f loral induction KP702299 namely MiAP1-1, MiAP1-2, MiAP1-3, and MiAP1-4. MiAP1-1 and MiAP1-2 are highly expressed in the f lowers, and overexpression in Arabidopsis significantly promotes f lowering [65]. MiAP1-2 has been associated with an early f lowering phenotype in transgenic tobacco [65]. It has been reported that MiLFY expression could be downregulated by exogenous gibberellin (GA3) and upregulated by paclobutrazol (PPP333). Bimolecular f luorescence complementation (BiFC) experiment showed that MiLFY protein could interact with zinc finger protein 4 (ZFP4) and CONSTANS overexpression inhibitor 1 (MiSOC1) to promote the early f lowering of A. thaliana [70]. Moreover, MiSOC1 has been isolated and identified from mango. Low ethephon concentration could upregulate MiSOC1 expression, but a high concentration inhibited MiSOC1 expression. Overexpression of MiSOC1 promoted the f lowering of A. thaliana [96]. SEPALLATA (SEP) gene has been reported to be highly expressed in mango inf lorescence [97]. Analysis of the distribution, phylogenetic relationship, subfamily division, gene amplification and evolution mechanism of gene families in the plant genome enables us to speculate on future gene evolution and function. Twenty-six SQUAMOSA promoter binding protein-like (SPL) family members were identified and analyzed in the 'SiJiMi' mango genome (unpublished data). Among them, 15 MiSPLs genes were highly upregulated during the early f lowering stage. Overexpression of MiSPL13 promoted the early f lowering of transgenic A. thaliana and the expression of AtAP1, AtSOC1, and AtFUL, which significantly improved the tolerance to drought, abscisic acid (ABA), and GA3 and was sensitive to Pro-Ca treatment [98].

Functional verification of mango fruit quality-related genes
Mango fruit ripening often starts at an early stage, and MiErpA1, MiCel1, MiERS1, MiETR1, and MiExpA1 play a role in fruit ripening and softening [72][73][74][75]99]. A study found that MiErpA1 expression was triggered within 90 min of ethylene treatment, and maturation-related transcription accumulation peaked on the third day after ethylene treatment. Importantly, 1-MCP treatment inhibited ripening/softening and MiExpA1 transcript and protein accumulation [72]. MiExpA1 expression may be ethylenedependent, and its expression increases during maturation. During maturation, the accumulation of MiCel1 transcripts gradually increases, related to increased EGase activity and decreased cellulose/hemicellulose content. The fruit ripening of control (ethylene treatment) and 1-MCP treatment was delayed by about 3 days, associated with a delayed increase in MiCel1 expression and EGase activity [73]. Oxalic acid significantly inhibited the decrease of pulp hardness and delayed MiExpA1 expression in peel and pulp. Oxalic acid alleviated cell wall disintegration during mango fruit storage, thus delaying the softening and ripening process of mango fruit [75]. Overwhelming evidence substantiates that MiETR1 and MiERS1 mRNA levels are upregulated with the prolongation of storage time, peaking on day 6. 1-MCP treatment significantly decreased MiETR1 expression on days 4, 6, and 10 and inhibited MiETR1 expression on days 2, 4, 6, and 10. These results indicate that MiETR1 and MiERS1 play an important role in ethylene signal transduction. The 1-MCP treatment effectively inhibited ethylene biosynthesis and ethylene-induced maturation and senescence [99].
The ripening and softening process of fruits involves the production and transport of cell wall polymers and enzymes. It has been established that Rab guanosine triphosphatases (GTPases) No. AY600964

MiACT1
Internal standard (Luo et al. 2013) [91] JF737036 are the main regulators directing traffic in the endomembrane systems. Twenty-three genes encoding RabA protein were identified using the existing mango transcriptome [21], and the relationship between pulp hardness and RabA gene expression of different mango varieties was studied, which substantiated the importance of pulp softening and transportation [82]. WRKY plays an important role in the plant defense regulatory network, development process and physiological processes such as various biotic and abiotic stress responses. Interestingly, 38 MiWRKYs genes correlated with mango malformation traits [100]. Compared with transcriptome analysis, searching the whole gene family through the genome can yield more comprehensive information. Polygalacturonase is a cell wall degrading enzyme that degrades pectin and participates in the softening of f leshy fruits during ripening. A total of 17 PGs cDNA were detected in the Kent mango peel transcriptome, while a total of 48 PGs genes were found in the mango genome, among which MiPG21-1, MiPG14, MiPG69-1, MiPG17, MiPG49, MiPG23-3, MiPG22-7, and MiPG16 were highly expressed during post-harvest fruit ripening, which may promote softening [38,101]. A total of 212 MibHLHs genes [102] and 315 MiWD40s genes were identified in the mango genome. Among them, MiTTG1 interacted physically with MiMYB0, MiTT8, and MibHLH1 in tobacco leaves [87], suggesting that a new ternary complex may be formed in mango, which may play an important role in plant defense regulatory network, development and other physiological processes.
With the development of metabonomics, secondary metabolites can undergo quantitative and qualitative analysis, emphasizing genes involved in metabolite synthesis. Current evidence suggests that Mi9LOX and MiEH2 participate in lipid biosynthesis. The concentrations of δ-valerolactone and γ -decalactone significantly increased when Mi9LOX was overexpressed, and the concentrations of δ-valerolactone, γ -hexalactone, and δ-hexalactone increased when MiEH2 was overexpressed, which further indicated that these genes might be involved in the biosynthesis of biogenesis of lactones from Alphonso mango [41]. It has been shown that the key structural genes of anthocyanin and proanthocyanidin synthesis in fruits, MiCHS, MiANS, and MiUFGT1, play an important role in the anthocyanin biosynthesis of mango peel. The MYB transcription factor regulates the expression of these genes. Compared with red mango (Guifei), green mango (Guiqi) and yellow mango (Jinghuang) produce fewer anthocyanins during maturity, secondary to the decrease in MiPAL activity at the translational level. It has been reported that the related transcription factors MiWRKY1, 3, 5, 81, and 84 are upregulated during light-induced anthocyanin accumulation, indicating that these genes may regulate the biosynthesis of anthocyanins in mango [77][78][79]. MiPSY, MiZDS, MiBCH, and MiZEP regulated the synthesis of carotenoids, and the transcripts were positively correlated with the total carotenoid content, but there was no significant difference in the expression of CRTISO among varieties. In addition, the differentially expressed carotenoid catabolism genes may explain the heterogeneity in carotenoid content among the three mango varieties. The expression of carotenoid catabolic genes (MiCCD1, MiNCED2, and MiNCED3) decreased faster in 'Kaituki', resulting in higher carotenoid content in 'Kaituki' than in the other two varieties [80,81]. However, no studies have hitherto reported on MYB transcription factors that specifically regulate the synthesis of polyphenol metabolites.
Few studies have been conducted on other tissue parts of the mango fruit. MiAUX1-4, an early auxin response gene, and MiPIN1, an auxin polar transport carrier element, promote root formation in Arabidopsis transgenic plants [103]. In 2017, Denisov et al. verified the response of MiLAX2 and MiPIN1 to growth. In A. thaliana, MiERS1a and MiERS1a responded positively to the ethylene signal [84]. However, the key enzyme genes MiACO and MiACS in the ethylene biosynthesis pathway and ethylene signal-related transcription factor MiERS1 could respond to salicylic acid and nitric oxide signals [85]. Without mango genome annotation, 18 complete MieIFs gene sequences were obtained through transcriptome data, and their expressions under salt stress, low-temperature stress and low-temperature stress were analysed. It was found that MieIF1A-a, MieIF5, and MieIF3sB might be candidate genes for improving the salt tolerance of mango [86]. Sixteen members of the Mi14-3-3 gene family were identified from the 'SiJiMi' mango genome database. By analysing their expression patterns under drought, salt stress and low-temperature stress, it was found that the Mi14-3-3 gene family played an important role under such stress conditions in mango [89].
At present, validation of gene function and the genetic transformation system of mango fruit have not been conducted due to low transformation efficiency and genotype dependence. Only some f lowering regulatory and carotenoid synthesis genes have been validated in plant models, including A. thaliana and tobacco. Accordingly, it is necessary to find an effective transformation system for mango.

Concluding remarks and future prospects
It benefited from the progress in sequencing technology, deciphering the mango genome and analysing fruit-related transcriptomes provide valuable data for variety identification, genetic diversity and transcriptional regulation of biological processes.
In particular, f lowering regulation, refining mango fruit appearance and growth characteristics has gained momentum. Future genome-wide association using multi-omics data, together with genetic population, natural population analysis and even accessions constructed pan-genomes will pinpoint regulatory genes for key traits, such as disease resistance genes, allowing easy and accurate breeding of new high-quality varieties in a short period of time, based on the focus on developing efficient mango tissue culture regeneration and gene editing technologies. These efforts may eventually bring a paradigm shift for mango breeding, which has substantial economic importance in tropical and subtropical regions.

Author contribution
M.S. drafted the original manuscript. Critical inputs and corrections were successively provided by H.W., Z.F., and H.H. during the preparation process. H.M. is the project leader and helped in the conception and structure design of the manuscript and final proofing of the manuscript for submission.

Data availability
Authors confirm the availability of data and that any required links or identifiers for data are present in the manuscript as described.