De novo assembly, gene annotation and marker development using illumina paired-end transcriptome sequences in nutmeg (Myristica fragrans Houtt.)

Abstract

Nutmeg (Myristica fragrans) is a high-value spice crop widely utilized in culinary, medicinal, and industrial sectors. Despite its economic and cultural significance, genetic research on nutmeg is severely limited, hindering efforts to improve its cultivation and breeding programs. The lack of molecular markers and genetic tools has further constrained the development of superior trait varieties. Molecular markers are essential for genetic mapping, diversity studies, and marker-assisted selection, yet their application in nutmeg remains unexplored. This study utilized Illumina high-throughput paired-end transcriptome sequencing to explore the genetic framework of nutmeg, focusing on de novo assembly, gene annotation, and molecular marker development. RNA Sequencing (RNA-seq) was performed on leaf tissues from male and female trees to analyze differentially expressed genes. A total of 141,192 unigenes were assembled with an average length of 1117 bp, achieving functional annotation in public databases like NR, Swiss-Prot, KEGG, and InterPro. De novo transcriptome assembly and functional annotation revealed ten candidate genes involved in sex determination. These candidate genes are associated with hormonal pathways, transcription factors, and reproductive development. Several primers/markers have been designed and validated using PCR. This research lays a foundation for genetic studies, marker-assisted breeding, and improved cultivation practices for nutmeg.

Introduction

Nutmeg belongs to the Myristicaceae family and is native to Indonesia, especially in East Indonesian regions such as the Moluccan Islands. (de Wilde 2000) states that the Myristicaceae family includes a large number of genera, with several species found in Indonesia, among which are Myristica fragrans Houtt., M. argentea (Warb.), M. fatua Houtt., M. gautterifolia DC., M. iners Bl., M. teysmanni Miq., Knema laurina (Bl.) Warb., K. cineria (Poir.) Warb. var., sumatrana (Miq.) Sincl., K. intermedia (Bl.) Warb., Horsfieldia glabra (Bl.) Warb., H. irya (Gaertn.) Warb., and H. iryaghedi (Gaertn.) Warb (Arrijani 2005). This circumstance allows nutmeg to develop further due to its richness in variations. At the same time, the mixed product is leading Indonesian nutmeg to have inferior quality so far.

Nutmeg is a dioecious evergreen tree that produces mace and seeds. One of the factors contributing to nutmeg’s low production is the conventional propagation method of utilizing seeds, which is still frequently used. Nutmeg’s dioecious nature causes genetic variance in seed-propagated plants, with a 55:40:5 chance of producing female: male: hermaphrodite trees (Flach 1966). This poses a significant barrier to nutmeg productivity because there is currently no accurate technology for early sex identification, which is essential for adjusting cultivation techniques for optimum productivity.

In recent decades, breakthroughs in high-throughput sequencing technology have enabled the generation of enormous amounts of transcriptomic and genomic data for a diverse range of model species. Transcriptomic data is the collection of all RNA molecules in a cell or tissue at a particular time, whereas genomic data is the entire DNA sequence of an organism’s genome. This data type can be used to identify genes expressed in specific tissues or under particular conditions, compare gene expression patterns across organisms, and examine the functional references between genes and their encoded proteins. The understanding of the genetic basis of complex traits has significantly increased due to the abundance of transcriptomic and genomic data available for these creatures. Gene prediction, gene annotation, gene ontology, gene expression level analysis, and identifying plant species’ metabolic pathway regulatory system may all be done using the vast amount of assembled transcriptomes (Kim et al. 2022). Even in non-model animals without a reference genome, RNA sequencing (RNA-Seq) provides a high-throughput method for acquiring vast amounts of transcriptome sequence data (Cao et al. 2018). Without a reference genome, RNA-Seq has been used to identify sequence variations (SNPs, SSRs) and create DNA markers. A powerful approach for transcriptome research, RNA-Seq is beneficial for crop sequence analysis without a reference genome sequence. Gene prediction, gene annotation, gene ontology, gene expression level analysis, and determining the regulatory system of metabolic pathways in plant species may all be done with the vast amount of assembled transcriptomes.

In this study, we used the Illumina high-throughput paired-end sequencing technology platform to analyze transcriptome data and to create a massive Expressed Sequence Tags (EST) dataset in several accessions of M. fragrans. These sequences will be useful for new gene finding, molecular marker creation, gene mapping, and comparative genomics.

Materials and methods

Plant material

Nine nutmeg accessions grown at the Indonesian Ministry of Agriculture’s experimental station in Cicurug, Sukabumi Regency, West Java, were assessed as material for RNA extraction and sequencing based on different origins and the tree sex types, male, female, or hermaphrodite trees (Table 1).

Table 1.

Nutmeg accession used as a sample for RNA extraction and sequencing

No	Sample code	Sample origin
1	1_MaleA	Banda-Male No. 1
2	2_FemaleA	Banda-Female No. 2
3	7_MaleA	Banda-Male No. 7
4	9_FemaleA	Banda-Female No. 9
5	32_FemaleA	Patani-Female No. 32
6	36_MaleA	Patani-Male No. 36
7	422_FemaleA	Ternate Irian Jati-Female No. 422
8	423_HermaprodhiteA	Ternate Irian Jati-Hermaprodhite No. 423
9	431_MaleA	Ternate Irian jati-Male No. 431

Male, female, and hermaphrodite trees were categorized based on the appearance and abundance of the flowering type of the tree (Fig. 1).

Fig. 1.

Female (a and b) and male (c and d) flowers of M. fragrans Houtt

Sample collection

Samples were collected from young leaves of the above nutmeg accession (Table 1), frozen immediately in liquid nitrogen, and stored at 80 °C until use. In some cases, it can scientifically justify the application of leaf samples for transcriptome analysis in sex determination, even though its effectiveness mostly depends on the plant species and whether sex-determining or sex-linked genes are expressed in that tissue. Several studies have successfully identified sex markers using leaf tissues through RNA-Seq-derived transcriptomic analysis, such as Coccinia grandis (Ghadge et al. 2014) and spinach (Bobadilla et al. 2023). To generate RNA-seq, shoot tip and internode tissues, instead of the reproductive organ, were applied in the transcriptome analysis of Salix purpurea L. (Carlson et al. 2017). (Akagi et al. 2014) also applied leaves to generate a genomic DNA library, bud tissues for mRNA, and a small RNA library of persimmons for sex determination. In this study, we applied young leaves due to low success in obtaining high-quality RNA for sequence analysis from flowers, flower buds, and root tissues (data not shown).

RNA extraction

Total RNA was extracted using the Trizol reagent according to the manufacturer’s instructions (Invitrogen). The total RNA concentration was quantified using UV spectrophotometry, and the quality of the total RNA was checked by electrophoresis in 1% agarose gel. Equal volumes of RNA from each of the four tissues were pooled. Paired-end libraries with approximate average insert lengths of 200 base pairs were synthesized using the Genomic Sample Prep kit (Illumina, San Diego, CA) according to the manufacturer’s instructions. Before cluster generation, library concentration and size were assayed using the Agilent DNA1000 kit (Agilent, USA) on a 2100 Bioanalyzer (Agilent, USA). Libraries were sequenced as 100-mer62 on a Hi-Seq 2000 equipped with a paired-end module at the Minin Gene Biotechnology Co. Ltd (Beijing, China).

Library preparation for transcriptome analysis

A total amount of 1.5 μg RNA was used as input material for the RNA sample preparations. The sequencing library was generated using NEBNcxt Ultra™ RNA Library Prep Kit for Illumina (NEB, USA). The mRNA was purified using poly-T oligo attached magnetic beads, and fragmentation was carried out using divalent cations under elevated temperature in NEBNext First Strand Synthesis Reaction Buffer (5 ×). First-strand cDNA was synthesized using a random hexamer primer and M-MuLV Reverse Transcriptase (RNase H-). Second-strand cDNA synthesis was subsequently performed using DNA Polymerase I and RNase-H. The remaining overhangs were converted into blunt ends via exonuclease/polymerase activities. After adenylation of 3′ ends of DNA fragments, the NEBNext Adaptor with hairpin loop structure was ligated to prepare for hybridization. The library fragments were purified with AMPure XP system (Beckman Coulter, Beverly, USA) insert size of 150–200 bp. Then, three μl USER Enzyme (NEB, USA) was used with size-selected, adaptor-ligated cDNA at 37 °C for 15 min followed by 5 min at 95 °C before PCR. At last, PCR products were purified, and library quality was assessed on the Agilent Bioanalyzer 2100 system.

Quality control, analysis, and assembly

Raw data (raw reads) of fasta format were firstly processed through in-house Perl scripts. The raw sequencing reads were filtered to generate quality reads for de novo assembly (Kang et al. 2017). This step obtained clean data (clean reads) by removing reads containing adapter, reads containing poly-N, and low-quality reads from raw data. At the same time, the clean data’s Q20, Q30, GC-content, and sequence duplication levels were calculated. All the downstream analyses were based on clean, high-quality data. The clean reads were de novo assembled with Trinity software, which is a high efficiency and stability transcriptome splicing software for RNA-Seq data, as well as cooperative research and development by Broad Institute & He- brew University of Jerusalem (Grabherr et al. 2011).

Illumina reads processing and assembly

A Perl script was written to remove low-quality sequences (reads with a base quality of less than 20). Then, the high-quality reads were assembled by Trinity with default settings except the K-mer value to construct unique consensus sequences.

Gene annotations and classifications

Functional annotations were performed by sequence comparison with public databases. All unigenes were compared with the NCBI non-redundant protein database (NR) (http://www.ncbi.nlm.nih. gov/), NCBI non-redundant nucleotide sequence database (NT), Swiss-Prot database (http://www.expasy.ch/sprot), the Clusters of Orthologous Groups database (http://www.ncbi.nlm. nih.gov/COG/) using Basic Local Alignment Search Tool (BLASTx) with different E-value of less than 1e-5 and 1e-10. InterPro domains were annotated by InterProScan Release 16.0, and functional assignments were mapped to Gene Ontology (GO) (http://www.geneontology.org/). Then we made GO classification and drew GO tree using WEGO (http://wego.genomics.org.cn/cgi- bin/wego/index.pl). Pathway assignments were carried out based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.genome.jp/kegg). Unigenes were first compared with the KEGG database using BLASTx with an E-value of less than 1e-10. Then a Perl script was developed to retrieve KO (KEGG Orthology) information from BLASTx results, and we established pathway correlation between unigenes and the database.

The bioinformatics analysis process starts by filtering raw sequence reads to remove low-quality reads (more than 20% of bases have a quality score of less than 10), adapter-contaminated reads, and reads with more than 5% unknown bases (N), leaving clean reads for further analysis. These clean reads are de novo assembled into transcripts using Trinity, which partitions the sequence data into de Bruijn graphs to reconstruct whole transcripts, then clustering with Toxoplasma gondii Isocitrate Lyase (Tgicl) to generate Unigenes. Unigenes are classified into clustered Unigenes, which have more than 70% sequence similarity, and single Unigenes. Functional annotation involves aligning Unigenes to databases such as NT, NR, KOG, KEGG, SwissProt, and InterPro using tools such as BLASTn, BLASTx, and InterProScan5. Blast2GO uses NR annotation to do Gene Ontology (GO) annotation. TransDecoder is used to predict candidate coding regions and identify the most extensive open reading frames (ORF), while BLAST and Hmmscan are used to identify protein homologs and conserved domains.

Primer design

Primers like MADS-box genes are developed for specific genes based on their function. These MADS-box genes, including blooming, fruit formation, and environmental responses, are essential in plant development. Primer3 is used to build primers, with requirements such as primer length (18–25 nucleotides), melting temperature (Tm) of 55–65 °C, GC content of 40–60%, and the avoidance of secondary structure development. After segregating male, female, and hermaphrodite groups, DNA sequences are screened for distinctness. All DNA sequences contain the Agamous gene, expressed throughout male and female flower development. The selected female group sequence comprises the Agamous gene and unique DNA sequences that differ from others. These sequences’ primers are designed using Primer3 and Primer Explorer v4, focusing on distinct variation areas for particular amplification.

Results and discussion

De novo transcriptome assembly

The de novo assembly by Trinity was performed on Clean Reads obtained in each sample by RNA-seq analysis, and 87,798–125,780 contigs were obtained. The average length of the contig was 728 to 816 bases. As a result of performing clustering with Tgicl on these contigs, 141,192 Unigenes were obtained. Since these Unigenes contain splicing variants derived from one gene and fragments of transcripts, they are judged to be larger than the actual number of nutmeg genes. RNA-seq analysis of RNA samples extracted from leaves of different sex types of nutmeg individuals resulted in 66.77–67.11 million Clean Reads generated after trimming the adaptor sequences, the ambiguous N nucleotides (> 0.1%), and the low-quality sequences. The ratio of Clean Reads to Raw Reads was 98.59–99.09%. Such a high ratio means RNA-seq analysis was performed commonly (Table 2).

Table 2.

Clean reads quality metrics

Sample	Total raw reads (M)	Total clean reads (M)	Total clean bases (Gb)	Clean Reads Q20 (%)	Clean reads Q30 (%)	Clean reads ratio (%)
1_MaleA	67.73	66.79	6.68	98.37	92.58	98.61
2_FemaleA	67.73	66.77	6.68	98.36	92.51	98.59
32_FemaleA	67.73	67.11	6.71	98.42	92.71	99.09
36_MaleA	67.73	66.79	6.68	98.42	92.7	98.62
422_FemaleA	67.73	67.01	6.7	98.44	92.71	98.94
423_HermaphroditeA	67.73	66.99	6.7	98.36	92.48	98.91
431_MaleA	67.73	66.83	6.68	98.4	92.62	98.68
7_MaleA	67.73	66.94	6.69	98.37	92.45	98.84
9_FemaleA	67.73	66.97	6.7	98.38	92.49	98.88

Clean reads must be assembled after sequencing for projects without reference genome to get a reference sequence for subsequent analysis. Then, after reading the filtered genes, we use Trinity to perform de novo assembly with clean reads; the assembly quality metrics are shown in Table 3. Tgicl was used on cluster transcripts to remove abundance and get Unigenes. The clustering quality metrics are shown in Table 4, and the Unigene length distribution is shown in Fig. 2.

Table 3.

Quality metrics of transcripts

Sample	Total number	Total length	Mean length	N50	N70	N90	GC (%)
x-1_MaleA	103,542	79,620,200	768	1,497	767	274	42.83
x-2_FemaleA	87,798	65,126,179	741	1,421	734	267	43.07
x-32_FemaleA	125,780	91,648,943	728	1,433	688	260	42.79
x-36_MaleA	108,497	85,399,324	787	1,564	806	278	42.33
x-422_FemaleA	121,808	94,589,090	776	1,542	777	275	42.36
x-423_HermaphroditeA	94,423	75,741,553	802	1,574	837	285	42.56
x-431_MaleA	91,325	74,575,048	816	1,588	865	290	42.69
x-7_MaleA	96,479	77,427,075	802	1,576	835	284	42.49
x-9_FemaleA	105,729	82,800,236	783	1,540	800	277	42.41

Table 4.

Quality metrics of unigenes

Sample	Total number	Total length	Mean length	N50	N70	N90	GC (%)
x-1_MaleA	75,686	68,014,483	898	1,633	924	331	42.92
x-2_FemaleA	65,032	56,295,099	865	1,553	889	320	43.19
x-32_FemaleA	90,601	78,022,357	861	1,589	862	313	42.82
x-36_MaleA	79,509	73,252,855	921	1,694	968	337	42.48
x-422_FemaleA	88,550	80,789,062	912	1,678	946	333	42.45
x-423_HermaphroditeA	70,157	65,372,958	931	1,692	987	343	42.70
x-431_MaleA	67,957	64,409,314	947	1,703	1,015	352	42.82
x-7_MaleA	71,682	66,625,831	929	1,692	982	341	42.62
x-9_FemaleA	77,995	71,538,423	917	1,679	961	335	42.56
All-Unigene	141,192	157,816,556	1,117	2,064	1,279	424	42.47

Fig. 2.

Unigene length distribution

Assembly evaluation and unigene functional annotation

The unigenes ranging in length from 300 to 3000 bp and ≤ 3000 bp showed the most significant numbers of unigenes at 300 bp (Fig. 2). The assembly-derived unigenes were searched against public protein and nucleotide databases for validation and annotation.

The summary statistics for annotation assignment against NT, NR, Swiss-Prot, KEGG, KOG, Interpro, and GO databases are shown in Table 5. As indicated, out of 141,192 unigenes, 66,870 (47.36%), 44,728 (31.68%), 47,132 (33.38%), 50,832(36.71%), 51,836 (36.71%), 58,178 (41,20%), and 40,213 (28.48%), unigenes showed annotation hits to homologous sequences in NT, NR, Swiss-Prot, KEGG, KOG, Interpro and GO databases, respectively. Besides the same, 44,728, 50,832, and 51,836 unigenes showed homology matches in the NT, KEGG, and KOG databases, respectively. In all, 19,295 (13.67%) assembled unigene sequences showed annotation hits to at least one of the protein and nucleotide databases, and 74,210 unigenes (52.56%) showed annotation hits to all databases. The reasons for unannotated sequences are several. Represent short sequences lacking a conserved protein domain. The transcriptome may also contain incompletely spliced introns, orphaned untranslated regions (UTRs), non-coding genes, and random transcriptional noise that may not show homology to available sequences in the public databases. An absence of annotations from the transcriptome would also mean genes expressed at low levels or not being expressed at the time of RNA extraction.

Table 5.

Annotation summary

Values	Total	Nr	Nt	SwissProt	KEGG	KOG	InterPro	GO	Intersection	Overall
Number	141,192	66,870	44,728	47,132	50,832	51,836	58,178	40,213	19,295	74,210
Percentage	100%	47.36%	31.68%	33.38%	36.00%	36.71%	41.20%	28.48%	13.67%	52.56%

A five-way Venn diagram plot (Fig. 3) also shows the representation of assembled unigenes annotated against InterPro, NR, KEGG, KOG, and SwissProt databases. This shows that all five databases concurrently annotated 35,421 unigenes, while 4835 unigenes had homologous sequences in Interpro, NR, KEGG, and KOG databases. Besides these unigenes, 3654 unigenes were annotated in the NR, KEGG, KOG, and Swissprot, and 4500 uni-genes were annotated in the interpro and NR databases.

Fig. 3.

Venn diagram between NR, KOG, KEGG, SwissProt, and InterPro

The E-value distribution indicated that 30.8% of annotated uni-genes have an E-value in the range of 0 to 1e-100, followed by 18.3% of unigenes with an E-value distribution of 1e-100 to 1e-60. The similarity distribution shows that 48% of the unigenes had a best-hit similarity of 60–80%, 32.1% of unigenes had a similarity of 40–60%, and 17.3% had a similarity of 80–95%.

KOG annotation

We annotate the Unigenes with the KOG database and calculate the Unigene distribution based on the 25 functional groups. Part of the KOG annotation results are shown in Table 6, and the functional distribution of KOG annotation is shown in Fig. 4.

Table 6.

KOG annotation result performance

Query_id	Subject_id	Align_length	E_value	Score
CL4092.Contig1_All	At1 g16190	394	2.6e-125	448.4
CL4092.Contig2_All	At1 g16190	394	2.8e-125	448.4
CL4092.Contig3_All	At1 g16190	394	2.3e-125	448.4
CL4093.Contig3_All	At4 g14210	282	1.9e-07	56.6
CL4094.Contig1_All	At5 g13780	180	1.7e-89	328.9

Fig. 4.

Functional distribution of KOG annotation

GO annotation

The Unigenes aligned to the NR database have been annotated to the GO database with Blast2GO, and we calculate the distribution in 3 categories: biological process, cellular component, and molecular functions. Unigene GO annotation results are shown in Table 7; the Functional distribution of GO annotation is shown in Fig. 5.

Table 7.

GO annotation result performance

Ontoloty	GO term	Number of genes	Gene members
Biological_process	Biological adhesion	7	CL2867.Contig2_All;…
Cellular_component	Cell	14,244	CL10003.Contig1_All;…
Molecular_function	Antioxidant activity	368	CL10077.Contig1_All;…

Fig. 5.

Functional distribution of GO annotation

DEG detection and KEGG

Nine RNA samples were subjected to analyse gene expression differences within and between classes based on genotype origin and sex (male/female/hermaphrodite). These resulted in 32 combinations, e.g., the same origin, different origin, or with hermaphrodite.

The selection of candidate genes for sex-specific molecular markers is carried out through a statistical analysis approach, starting with the identification of genes that have a False Discovery Rate (FDR) value of zero. This stringent criterion ensures high confidence in the selected genes by minimizing the likelihood of false positives. Then, transcription factors associated with flowering regulation and sex determination are identified and analysed. This step is crucial to selecting candidate genes with strong biological relevance to sexual differentiation mechanisms in plants. The ultimate goal is to develop reliable and precise molecular markers that can be effectively utilized in breeding programs for early sex identification, thereby improving the efficiency of cultivar development and selection processes of nutmeg (Table 8). The number of these genes (DEGs) is shown in Fig. 6.

Table 8.

Differential gene expression with a comparative combination between male and female tree (Up-regulated genes involved in reproductive systems in female tree compared to male)

Gene ID	Length	Male expression	Female expression	Description
CL8208.Contig3_All	2328	10.79	61.06	XP_021620764.1/5.9e-259/auxin transporter-like protein 2 [Manihot esculenta]
CL8035.Contig1_All	904	80.94	280.18	AEW24432.1/5.1e-17/late embryogenesis abundant protein 3L-1 [Camellia sinensis]
CL12054.Contig1_All	1387	10.89	200.45	OVA06732.1/1.8e-90/C2 calcium-dependent membrane targeting [Macleaya cordata]
CL2115.Contig1_All	2545	0.01	62.98	OVA13014.1/1.0e-123/AP2/ERF domain [Macleaya cordata]
Unigene21392_All	1655	0.01	42.78	OVA13014.1/1.0e-123/AP2/ERF domain [Macleaya cordata]
CL7815.Contig2_All	908	193.97	1128.39	OVA18739.1/1.5e-29/SANT/Myb domain [Macleaya cordata]
Unigene33873_All	2024	0.67	36.86	AFH03054.1/7.1e-152/R2R3-MYB transcription factor MYB2 [Epimedium sagittatum]
CL2746.Contig2_All	2744	15.93	63.68	XP_011629028.1/6.7e-137/two-component response regulator-like APRR9 [Amborella trichopoda]
CL7060.Contig2_All	2121	0.09	79.57	OVA19634.1/3.0e-233/Nodulin-like [Macleaya cordata]
Unigene28911_All	1774	5.1	51.66	EOY15860.1/1.6e-94/PLATZ transcription factor family protein isoform 3, partial [Theobroma cacao]

Fig. 6.

The distribution of gene expression across nine RNA samples of nutmeg from different origins. The X-axis represents the sample names, while the Y-axis represents the number of genes. The color intensity indicates different levels of gene expression: FPKM ≤ 1 (light blue) denotes very low expression genes; FPKM 1–10 (medium blue) denotes moderately expressed genes, and FPKM ≥ 10 (darkest blue) indicates high expression genes

Sex determining gene always locates in the sex chromosome. In our study, 87,138 of approximately 720,000 sex-chromosomal genes showed high expression level (Fig. 6). Moreover, out of those genes, there were 10 candidate genes which might be played as sex determine-genes in M. fragrans (Table 8). These candidate genes are associated with a various biological functions, including auxin transporter-like protein 2 (Balzan et al. 2014), late embryogenesis abundant protein 3L-1 (Manfre et al. 2006), C2 calcium-dependent membrane targeting (Sun et al. 2021), AP2/ERF domain (Wang et al. 2023),/SANT/Myb domain (Zhang et al. 2024), R2R3-MYB transcription factor MYB2 (Zhang et al. 2024), two-component response regulator-like APRR9 (Hwang et al. 2002), Nodulin-like (Hou et al. 2016), and PLATZ transcription factor family protein isoform 3, partial (Han et al. 2022). These findings provide valuable insights into the genetic mechanisms underlying sex determination in M. fragrans.

Genes labeled up or down have significantly different expression levels between the samples (DEGs). Due to the large number of DEGs detected among the samples, Gene Ontology analysis and Pathway analysis were performed to identify the characteristics of each gene. These results are presented in 10 Gene Ontology Analysis of DEGs (Fig. 7) and 11 Pathway Analysis of DEGs (Fig. 8).

Fig. 7.

Gene Ontology (GO) classification of differential gene expressions

Fig. 8.

Pathway classification of DEG

GO encompasses three ontologies: molecular function, cellular component, and biological process. A pathway analysis was performed to identify the characteristics of DEGs (Fig. 9). Based on the Pathway analysis, the characteristics of DEGs can be understood, and some speculations about the mechanisms underlying phenotypic differences between the samples can be drawn.

Fig. 9.

Functional distribution of KEGG annotation

The X-axis represents the number of DEGs, while the Y-axis represents the KEGG functional classification. KEGG pathways have seven branches: Cellular Processes, Environmental Information Processing, Genetic Information Processing, Human Diseases (only for animals), Metabolism, Organismal Systems, and Drug Development. However, the results indicate that some DEGs detected are classified as animal/human pathways.

Annotate the Unigenes with the KEGG database and calculate the Unigene distribution of KEGG Level 1 and KEGG Level 2, respectively. The functional distribution of KEGG annotation results is shown in Fig. 9. Part of the KEGG annotation results are shown in Table 9.

Table 9.

KEGG annotation result

Pathway_level1	Pathway_level2	Number_of_genes	Genes
Cellular processes	Transport and catabolism	2,286	CL10008.Contig2_All;…
Environmental information processing	Membrane transport	542	CL104.Contig10_All;…
Genetic information processing	Folding, sorting and degradation	3,150	CL10008.Contig1_All;…
Metabolism	Amino acid metabolism	2,867	CL10070.Contig1_All;…

Sarmah et al. 2024 revealed significant findings on Myristica fragrans’ bZIP transcription factors, which are essential for plant defence mechanisms, by the transcriptome analysis data. The study found 15 bZIP transcription factors highlighted by Gene Ontology (GO) analysis and the functions linked to defence responses and abscisic acid signalling pathways, suggesting they are involved in plant stress responses (Sarmah et al. 2024). According to KEGG Pathway Analysis, the MfbZIP factors have binding motifs for proteins involved in hormone signal transduction, which is essential for disease resistance (Sarmah et al. 2024). However, there was no information related to transcription factors associated with sex determination. The same result was also revealed in other studies on Myristica fatua, in which its de novo transcriptome data merely lacked datasets related to sex-determination genes (Matra et al. 2023).

Unigene TF prediction

After assembly, we predict Unigenes that encode Transcription Factor (TF) for plant research. The list of Unigenes that encode TF in our project is shown in Table 10, and we also conduct classification on the TF family, shown in Fig. 10.

Table 10.

TF prediction result

Gene_id	TF_family	Included_domain	Excluded_domain
CL2783.Contig4_All	MADS	SRF-TF	–
Unigene48570_All	Tify	tify	GATA
CL584.Contig6_All	HSF	HSF_DNA-bind	–
CL2469.Contig4_All	BES1	DUF822	–

Fig. 10.

Classification of transcription family

Primer design

Trinity software used the de novo assembly method to analyze and clarify the sequencing reads from each leaf sample. The results indicated different contigs, ranging from 87,798 to 125,780, with an average contig length of 728 to 816 bases.

The TGICL (TIGR Gene Indices Clustering Tool) algorithm performed further clustering analysis, identifying 141,192 Unigenes. This clustering provides a thorough understanding of gene expression in nutmeg leaf samples. Once gene expression was mapped, additional research was performed to categorize the biological roles of the expressed genes, including the identification of sex genes.

Sex-related gene identification was carried out using PCR and sequencing, which resulted in finding specific genes expressed in male and female plants. Identifying 33 flowering-associated genes, classified as members of the MCM1, AGAMOUS, DEFICIENS, and SRF (MADS)-box gene family, was a significant discovery. These genes play an essential role in floral development and may provide insights into the genetic mechanisms underlying sex differentiation in Myristica fragrans.

The transcriptomic analysis of 9 different individual nutmeg trees, originating from various locations and of different sexes, based on the characteristic of fruiting/non-fruiting, using 23 combinations through Next Generation Sequencer (NGS), detected no significant differences in gene expression among individuals of the same sex. This means that gene expression is identical for the same sex. Subsequently, DNA-specific sequences were adopted and designed for Loop-Mediated Isothermal Amplification (LAMP) primers and confirmed by RT-PCR. Ten LAMP primers were designed and applied to amplify DNA from male and female nutmeg trees. Three primers among 10 (Table 11), namely Mf1, Mf3, and Mf7, consistently produced band patterns that appeared in some male nutmeg samples and were absent in female samples.

Table 11.

Ten LAMP primers designed from DNA-specific sequence adopted from the de novo sequence of the nutmeg genome

No	Primer name	Sequences
		Forward	Reverse
1	LAMP 1	CGTTGAAGGAGTCGCAGC	CCGGTATGGTCGTCACCTT
2	LAMP 2	TTGGGGAACAGGAAGTTTTG	TCAGAGCTTCAAAAGCACCA
3	LAMP 3	GCAGGATGGGAAGGGTTAAG	TTCTCAAAATTTTCCTTATG
4	LAMP 4	AGAAGAGGATCAATGCGGGC	AGCCTGATTGTTGCTGCATT
5	LAMP 5	AGTTTGCAGGATGGGAAGGG	GGCCTGACTAGAGTTGACTCC
6	LAMP 6	AGCCCATGTCGTGGTTTCAT	GCCCACTAACCTTGGAGTCA
7	LAMP 7	TGATTTTGCAGGGGAGGTGT	AAAGCACCAGAACAAGGGGT
8	Mf 1	AAGGCCAGGGAGCTATCAAT	AGTGCCTCAAGGCTCTCCAG
9	Mf 2	AAGGCCAGGGAGCTATCAAT	GTGCCTCAAGGCTCTCCAGT
10	Mf 3	AAGAAGGCCAGGGAGCTATC	AGTGCCTCAAGGCTCTCCAG

Although some deviations still occurred, this finding indicates that the designed LAMP primers could be developed as markers for determining the male sex of Myristica fragrans (Fig. 11). There were no significant differences amongst the three primers-LAMP applied to 12 individuals’ male/female tree DNA (according to 0.05 Chi² test, data not shown). To further verify the ability of these primers to detect the sex of nutmeg at an earlier stage, an in-depth study is needed by applying samples from a wider population.

Fig. 11.

Amplification products of LAMP primers Mf1 (a), Mf 2 (b) and Mf3 (c) on Nutmeg DNA from 12 individuals of different sexes, 6 males and 6 females, produced a specific banding pattern for the male sex (arrowed). M = DNA ladder 100 bp, 1 = male, 2 = female, 3–6 = male, 7–9 = female, 10 = male, 11–12 = female

One of the limitations of de novo sequencing is that the fragmentation and DNA sequence reading can only achieve up to 40%, resulting in shorter readable sequences compared to those with a reference genome library. The accuracy of de novo sequencing in plant genomes is influenced by several critical factors, including sequencing technology, base calling methods, and the genomic architecture of the target species. Understanding these elements is essential for improving assembly quality and accuracy. Long-read sequencing technology enhances contiguity and accuracy in assembly and can also show and achieve near-complete chromosome-level assemblies (Sharma et al. 2022). Further, Longer-reads are more effective in resolving complex genomic features, such as repetitive regions, which are challenging for short-read technologies (Bucur 2017).

Meanwhile, plant species-specific base caller improves the read accuracy by up to 96.87% (Ferguson et al. 2022). Due to increased genetic variety and repetitive sections, many plant species have significant levels of heterozygosity and polyploidy, which makes assembly more difficult (Zhou et al. 2022). With a chromosomal number of 2n = 2x = 42, Myristica fragrans is a diploid species that does not display polyploidy (Nair 2019). However, research indicates that nutmeg has a significant degree of heterozygosity. In one investigation, for example, the observed heterozygosity across loci varied from 0.038 to 0.929 for 48 nuclear microsatellite markers. An average heterozygosity of 1.00 was reported by another study that included 17 SSR markers, indicating substantial genetic variation within the species (Kusuma et al. 2020). Therefore, further research on the genome sequence of nutmeg is essential to explore.

In papaya, the trioecious species, de novo sequencing has revealed significant insights into the genetic basis of sex differences. Female plants have XX chromosomes; male and hermaphrodites have an XY. The sequencing effort has mainly focused on the specific region of the Y chromosome (HSY) in hermaphrodite (VanBuren and Ming 2014). Hermaphrodite sexes are the focus of research on papaya due to the advantages in larger, smooth, and fleshy fruits compared to females and being able to self-pollinate, resulting in consistent high yield. Therefore, sex determination in papaya plants by identifying and planting only hermaphrodite plants offers several advantages (Guadalupe Ávila-Hernández and Martinez-Antonio 2023). From field observation, unlike papaya, hermaphrodite Indonesian nutmeg mainly produces smaller fruit than the female tree, which is pollinated by flowers derived from the male tree. However, the more male trees in the field, the fewer chances of land productivity being observed. Therefore, sex determination in nutmeg becomes essential to differentiate male trees from female ones significantly.

Conclusion

This study provides a significant understanding of the differentiation between male and female Myristica fragrans Houtt. through molecular and transcriptomic analysis. The results revealed distinct gene expression profiles associated with hormonal pathway, transcription factor, and reproductive development, highlighting crucial regulatory mechanisms in sex determination. Besides that, these finding became the potential of molecular breeding for early sex determination, which is essential for optimizing nutmeg cultivation and improving yields. The integration of genomic and transcriptomic data enhances our understanding of the mechanisms regulating sex determination in dioecious species, providing a foundation for future breeding program. The findings also emphasize the potential for further research into the genetic basis of nutmeg traits, paving the way for advanced breeding strategies and conservation efforts.

Author contributions

OR: designed and prepared genetic materials, evaluated the accessions, observed and analyzed data, and wrote the manuscripts; RH and TPP: designed, analyzed data, designed primer, and wrote the manuscripts; TA, AR, and MRH: designed and prepared genetic materials, and wrote the manuscripts. This article was reviewed and authorized by all authors.

Funding

Kementerian Pertanian Republik Indonesia, Badan Riset dan Inovasi Nasional.

Data availability statement

The data supporting the findings of this study will be made available upon request to the corresponding author.

Declarations

Conflict of interest

The authors declare no conflict of interest.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.

Consent for publication

All the authors have read and consented to submit the manuscript.