Palmae is a monocot family containing 202 genera and approximately 2800 species mainly distributed in tropical and subtropical regions. Palm trees generally have a solitary shoot bearing a crop of very large leaves. The majority of palm trees are used as landscape plants, planted along both sides of the street. However, there are also some important economic crops in this family, such as Elaeis guineensis, Phoenix dactylifera, Cocos nucifera and Areca catechu. Of the four economic crops, the African oil palm E. guineensis has the largest plantation area (17 million hectares), producing 50 million tons of palm oil annually. Due to its economic importance, the whole genome sequence of E. guineensis has already been completed and was released in August 2013 (Singh et al., 2013). The draft genome sequence of related species P. dactylifera has also been reported (AI-Dous et al., 2011; AI-Mssallem et al., 2013). The genome sizes of E. guineensis and P. dactylifera are 1.535 and 605 Mb, respectively. However, only 657.86 Mb (42.86%) can be mapped into the 16 assembled chromosomes of E. guineensis, and some chromosomes are not assembled in P. dactylifera. Previous research based on phylogenetic dating has shown a divergence 65 Mya between E. guineensis and P. dactylifera (Singh et al., 2013). Except for these species, other Palmae are still under-studied and few genetic and genomic resources are available. This is an obstacle to the conservation and genetic assessment of Palmae species, especially those that are currently endangered.

Microsatellites (simple sequence repeats, SSRs) are tandem DNA repeats which are mostly distributed throughout non-coding regions of eukaryotic genomes. Microsatellite markers are generally co-dominant and high polymorphic, and are widely used for constructing genetic maps, diversity analysis and quantitative trait locus mapping (Qiu et al., 2006; Xiao et al., 2010; Zhang et al., 2012; Xia et al., 2014). Due to lack of selection pressure on these non-coding sequences, microsatellites are generally regarded as neutral genetic markers. Recently, microsatellites have been also validated to have important biological functions, such as the regulation of chromatin organization, DNA metabolic processes, gene activity and RNA structure (Li et al., 2002, 2004; Gemayel et al., 2010). Some research has already been done to identify microsatellite markers via probe hybridization against genomic/cDNA clones (Rivera et al., 1999). This method of developing microsatellites is still low efficiency and high cost. With the use of next-generation sequencing (NGS) technology, it became possible to develop large numbers of molecular markers through in silico analysis of available bacterial artificial chromosome (BAC) sequences (Xu et al., 2010), BAC-end sequences (BES) (Bohra et al., 2011), genome survey sequences (WGSs) (Marek et al., 2001), expressed sequence tags (Low et al., 2008), and whole genome shotgun sequences (WGSs) (Shi et al., 2013). In the Palmae family, cDNA library tags and transcriptome data have been used to develop EST-SSRs in E. guineensis (Xiao et al., 2014), P. dactylifera (Zhao et al., 2012), and C. nucifera (Xiao et al., 2014). With the release of the whole genome sequences of E. guineensis and P. dactylifera, this will provide an opportunity to develop larger number of microsatellite markers and ascertain their distribution across different chromosomes in the two Palmae species. However, except for these two Palmae species, identification of microsatellite markers is difficult because of the lack or scarcity of available sequence data in other Palmae species.

It has been well recognized that some molecular markers (including microsatellite markers) can be transferable from different genotypes within or between species or even between genera (Wang et al., 2012). Such interspecific or intergeneric transferability for markers will provide some useful genetic information for poorly-studied related species, contributing to conservation, genetic assessment, and construction of linkage maps. Meanwhile, the transferability of microsatellite markers depends on the genetic relationship between the species. Mnejja et al. (2010) reported that microsatellite markers developed from peach and almond had a higher transferability rate to other Prunus species than to three other non-Prunus rosaceous genera (Mnejja et al., 2005). In previous research, microsatellite cross-species transferability is identified generally via PCR amplification in different related species using primer pairs designed from one species (Decroocq et al., 2003; Gasic et al., 2009). This method to identify cross-species transferability of primer pairs is very low efficiency. Meanwhile, the number of primer pairs with cross-species transferability identified is also limited.

Among the Palmae family, the whole genome sequences of E. guineensis and P. dactylifera have been completed and released. This study (1) identified a large number of microsatellites and analyzed the frequency and distribution of microsatellites in the two palm species, respectively; (2) analyzed the distribution of these microsatellites in 16 chromosomes of E. guineensis; (3) identified conversed microsatellites genome-wide between E. guineensis and P. dactylifera; (4) validated the cross-genera transferability of these conserved microsatellites in Palmae.

Although the genomic sequences of E. guineensis and P. dactylifera have been released, there are still no related documents which elucidate and compare the frequency and distribution of microsatellites in the two palm species. To our knowledge, this study is the first report on genome-wide identification and comparison of microsatellite loci based on the assembled genomes of E. guineensis and P. dactylifera. In previous studies, microsatellites were generally developed based on the released transcriptome data from E. guineensis and P. dactylifera. Tranbarger et al. (2012) detected 465 microsatellites within 2 652 262 bp of E. guineensis EST sequences. Zhao et al. (2012) detected 5981 microsatellites from 28 889 assembled EST sequences from P. dactylifera. However, our study was based on the entire assembled genomes of E. guineensis and P. dactylifera, including non-coding sequences. This allowed a far greater number of microsatellites (814 383 and 371 629 respectively) to be identified compared to previous studies (Tranbarger et al., 2012; Zhao et al., 2012; Aberlenc-Bertossi et al., 2014).

Previously, the frequency of microsatellites was reported to be very similar in different species of the same genera (Shi et al., 2013). In this study, E. guineensis and P. dactylifera are classified into different genera in the Palmae family. However, the density of microsatellites in the assembled genomic sequences of E. guineensis (770.4 microsatellites per Mb) and P. dactylifera (733 microsatellites per Mb) is also nearly identical, indicating highly similar genomic structure in the two palm species. Meanwhile, the distribution with respect to the motif type and repeat number of microsatellites in the assembled genomic sequences of the two palm species is also very similar. More specifically, A/T rich motifs are generally dominant/major, while C/G rich motifs are absent/scarce, which is in accordance with reported results by Shi et al. (2013). Microsatellite expansion and variation is thought to result from replication slippage by mismatch during DNA replication (Tautz and Schlotterer, 1994; Klintschar et al., 2004; Forster et al., 2015), which occurs more readily in A/T rich motifs than in C/G rich motifs. Overall, the microsatellite distribution characteristics in the assembled genomics of the two palm species are very similar. This suggests that the two Palmae species have similar genome structure, which may be due to their relatively recent divergence (65 Myr) from a common ancestor (Singh et al., 2013).

In this study, a total of 814 383 and 371 629 mono- to hexanucleotide repeat microsatellites were identified in the assembled genomic sequences of E. guineensis and P. dactylifera, respectively, with frequencies of 770.4 and 733 microsatellites per Mb or one microsatellite per 1.29 and 1.36. The microsatellite density in the assembled genomic sequences of the two Palmae species is higher than the SSR density in some species (Tautz and Schlotterer, 1994; Klintschar et al., 2004; Song et al., 2015). To our knowledge, this study is the first report on genome-wide identification of microsatellite loci in palm family. Most (99.29 and 98.39% for E. guineensis and P. dactylifera) of the identified microsatellite loci are novel compared to previous research in these two palm species (Zhao et al., 2012; Xiao et al., 2014). Although further validation may still be required, these newly developed microsatellite loci will be useful for future molecular marker studies in E. guineensis and P. dactylifera. Possible applications include the construction of high-density linkage maps, QTL fine mapping and genome-wide association mapping.

A great deal of research has been performed to identify genic microsatellites via in silico analysis of publicly available unique transcriptome sequences in E. guineensis (Tranbarger et al., 2012; Xiao et al., 2014). However, due to the lack of genomic sequences, the distribution and physical position of these publicly available microsatellite markers is still unclear, which has a negative effect on their effective utilization. In this study, we describe the genome distribution and the accurate physical position of these newly developed microsatellite markers based on the mapped sequence scaffolds of E. guineensis. The high-density microsatellite markers with accurate physical position could be very useful for rapid selection of trait-associated markers and subsequent application of molecular breeding in E. guineensis. Previous studies have showed that the distribution of microsatellites is preferentially associated with non-repetitive DNA/gene sequences in plant genomes (Morgante et al., 2002; Gemayel et al., 2010). In our study, the genomic distribution of microsatellites was also positively correlated to that of genes but negatively correlated to that of TEs. The high accordance between the genomic distribution of microsatellites and genes may suggest a putative role of microsatellite in regulating gene function (Li et al., 2002, 2004; Gemayel et al., 2010). Some microsatellite markers could also be used as functional markers to tag genes and for molecular breeding.

Microsatellite markers obtained from one plant species are seldom used directly in populations of closely related species. In order to screen cross-genera transferability of microsatellite markers, primer pairs flanking microsatellites from one plant species are generally used to amplify DNA from other species in the same genera (Chagne et al., 2004; Guo et al., 2006). This method has comparatively low efficiency. Normally, only a few cross-species transferable microsatellite markers can be obtained through the selection of a large number of primer pairs. This study was conducted to select cross-genera-transferable microsatellite markers in the Palmae family. In order to increase the selection efficiency, a large number of conserved microsatellites were first identified between E. guineensis and P. dactylifera. Most of these conserved microsatellites identified in the present study were also validated to have high transferability ratio in palm species in the study. In conclusion, the conserved microsatellites identified in the study will provide some genetic resources for the assessment of genetic diversity, the analysis of population structure, and conservation of palm species, especially for endangered species.

Electronic PCR (e-PCR) is a useful tool for simulating the process of Polymerase Chain Reaction (PCR) and for detecting potential amplification products. In the present study, validating transferability for 618 762 primer pairs would be prohibitively time-consuming. Electronic PCR (e-PCR) simplified the process. The selection of parameters is important for the reliability of electronic PCR (e-PCR). The default parameters of the software are 2 bp mismatch, 1 bp gap, and 500–1000 bp product size. In the present study, in order to increase the reliability of the electronic PCR results, more stringent parameters were applied: a 1 bp mismatch and 0 bp gap were allowed in order to decrease the chance of false positives, although this may also increase the false negative rate.

Location of microsatellites in non-coding regions allows easier accumulation of mutations over the generations, and gives rise to increased variability which can be used for DNA markers. However, other microsatellites may be located in the UTR of mRNA, promoter regions of expressed genes, or even directly in CDS of genes. In such cases, microsatellite mutation may lead to phenotypic changes. Microsatellite variation in CDS can lead directly to functional protein changes, while microsatellite variation occurring in 5′ UTR (5′-UTRs) may affect transcription and translation, and microsatellite variation in 3′-UTRs may affect splicing. Therefore, these microsatellites associated with biological functions due to their location may be more highly conserved between different genotypes and even between different species, as most novel mutations are detrimental. In this study, some microsatellites were found to be conserved between palm species, and the position of these conserved microsatellites in the genome of E. guineensis was also analyzed. The majority of the conserved microsatellites (71.45%) were located within genic regions, while the remainder (28.55%) were located within intergenic regions. It is possible that some conserved microsatellites in intergenic regions were located on promoter regions of genes, such that mutation of these microsatellites may lead to a change in gene expression. Mutation of the 71.45% of conserved microsatellites located in genic regions may directly lead to changes in amino sequence, splicing, and gene expression. In summary, microsatellites associated with biological function due to their genomic location in genic regions may be more conserved between different genotypes and even species, and hence more transferable to other species. Therefore, selection of genic microsatellites may be a useful strategy for identifying high proportions of markers transferable between species.

In the last two decades, advances in next-generation sequencing technology have facilitated the development of molecular markers. In this study, genome-wide identification of microsatellites was conducted based on the assembled genomic sequences of E. guineensis and P. dactylifera. A high density of microsatellites was identified in the two species, which will provide a useful tool for research into genetic diversity, gene mapping and molecular breeding in these species.

It has been previously documented that some microsatellites could be transferable between species or even between genera (Chagne et al., 2004; Mnejja et al., 2005; Guo et al., 2006; Pandey et al., 2013), In this study, we tried to develop a large number of conserved microsatellites which can be used for genetic analyses in different palm species. A total of 7 265 conserved microsatellite markers were identified between E. guineensis and P. dactylifera. In the future, these identified conserved SSR markers may be regarded as useful genetic resources for diversity conservation in different species in the Palmae family, especially in endangered species.

YX and WX did the DNA extract and subsequently PCR amplification, participated in the design of the study, performed the statistical analysis and drafted the manuscript. JM did the extraction of DNAs and PCR amplification. AM critically revised the manuscript. XL, HF, ZM, and MP participated in the design of the study. All authors read the final manuscript and have given final approval of the version to be published.

This work was supported by the Natural Science Foundation of China (No.31301358), the Scientific and Technological Cooperation Projects of Hainan province (No.SQ2015GJXM0076), National Nonprofit Institute Research Grant of CATAS-ITBB (No.1630052015050), and The Major Technology Project of Hainan (ZDZX2013023-1).