An enhanced gene function analysis platform for Gastrodia Elata integrating further omics data and analysis tools

Abstract

Gastrodia elata, the dried tuber of the Gastrodia orchid, is a valued traditional Chinese medicine with important medicinal and dietary benefits. The growing multi-omics data demands an updated platform to include the latest G.elata genome and transcriptome, thereby enhancing gene function analysis for researchers. In this study, we built co-expression network using four versions of the G.elata genome, incorporated the latest transcriptome data, and annotated genes with sequences from TAIR, Uniprot, NR, and Swisprot databases. GO and KEGG annotations were predicted using InterProScan and GhostKOALA, while gene families were identified with HMMER, iTAK, OrthoFinder, and KEGG annotations. We further completed the GelFAP v3.0 upgrade (http://www.gzybioinformatics.cn/Gelv3), integrating results and adding tools like BLAST, Gene sets enrichment analysis, Extract sequence, JBrowse v2, multiple sequence alignment, DEG analysis, DEM analysis, primer design and heatmap analysis. Through this platform, we identified that the GeMYB4 gene may function as a potential suppressor of flavonoid biosynthesis. Additionally, the platform enabled the identification of GePAL1, a key gene in the phenylpropanoid pathway, which showed higher expression in G.elata f.glauca compared to f.elata and was co-expressed with other upregulated genes involved in phenolic compound biosynthesis. Using GeMYB4 and GePAL1 as cases, we demonstrate the utility of this platform. GelFAP v3.0 will aid in analyzing key genes related to the biosynthesis of secondary metabolites, stress response, growth and development, offering researchers enhanced resources for studying G.elata gene functions. 

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-025-07173-7.

Introduction

Gastrodia elata, or Tianma, is a medicinal plant used in traditional Chinese medicine for its pharmacological benefits. G.elata extracts have shown to reduce cognitive impairment and enhance learning and memory in Alzheimer’s model mice by affecting neurotransmitter metabolism [1, 2]. Additionally, its water extract shows promise in reducing depressive symptoms from chronic mild stress, indicating potential mental health applications [3]. The bioactive components of G.elata, including phenolic compounds and polysaccharides, boost its medicinal value by offering antioxidant and anti-inflammatory benefits and potentially improving metabolic disorders through gut microbiota modulation [4, 5]. As a traditional Chinese medicinal herb, G.elata shows significant therapeutic potential, meriting further study in modern medical research. Its main component, gastrodin, has notable neuroprotective properties that help relieve symptoms of neurological disorders like epilepsy, headaches, and dizziness [6, 7].

The genome sequencing of G.elata has revealed key genetic insights. Bae and team produced a high-quality chromosome-level genome assembly of this orchid, highlighting its reliance on mycorrhizal fungi and evolutionary adaptations to a mycoheterotrophic lifestyle [8]. Yuan’s team sequenced the 1.06 Gb genome of G.elata, finding 18,969 protein-coding genes and significant gene loss, including photosynthesis genes [9]. Xu’s team created a chromosome-level genome of G.elata and compared it with other orchids and plants [10]. Wang’s team assembled a 1.04 Gb draft genome of G.elata, identifying three horizontal gene transfer events among 17,895 protein-coding genes [11]. Researchers analyzing the G.elata genome have identified genes linked to its pharmacological effects and their roles in biosynthetic pathways. Transcriptomic studies offer insights into gene expression regulation, revealing genes that are upregulated or downregulated during different growth stages and conditions, which are crucial to G.elata’s growth, development, and pharmacological activity [12, 15].

We previously created two versions of the G.elata gene function analysis platform, each using data from just one genome assembly [16, 17]. However, current genomic data remains fragmented and lacks systematic integration, hindering data sharing and limiting a comprehensive understanding of the genetic traits of G.elata. Furthermore, integrating and analyzing transcriptomic data remains challenging, particularly in combining it effectively with other omics data. To address these issues, this study established a systematic multi-omics platform that integrates genomic, transcriptomic and metabolomic data of G.elata, promoting data sharing and collaborative research. Additionally, the development of new analytical tools and methods will further facilitate the discovery of genes related to the biological characteristics of G.elata.

Materials and methods

Data resources

Genomic data for G.elata were sourced from the National Genomics Data Center in China (NGDC), 89 transcriptome samples in this study were downloaded from Short Read Archive (SRA) database and 6 samples was produced by our group (Table S1). TrEMBL and SwissProt sequences from theUniprot database [18], TAIR protein sequences were downloaded from the TAIR website [19], and NR protein data were obtained from the non-redundant protein sequence database in the NCBI. KEGG annotation information was collected in the KEGG database, and GO annotation information was obtained from the Gene Ontology Consortium [20]. The carbohydrate-active enzyme (CAZy) protein sequence is derived from the CAZy database [21], and the Pfam domain annotation is derived from the InterProScan database [22].

Functional annotation

The protein sequences of G.elata were analyzed using Diamond BLAST [23] by comparing them against public databases, including SwissProt, trEMBL, NR, and TAIR. The best-matching results from these databases were then selected and compiled to provide the annotated information. KEGG annotation was generated using the GhostKOALA online tool [24]. GO annotation and Pfam domain information were predicted using InterProScan software [22].

Gene family identification

The orthologous protein relationships between A.thaliana and G.elata were predicted using the OrthoFinder software [25]. Utilizing the orthologous relationship, we identified the proteins associated with CAZy and TP. The KEGG pathway of the G.elata genome was annotated with GhostKOALA [24], and CYP450 gene functions were determined from these results. In addition, we obtained the hidden Markov model from iUUCD v2.0 [26] to complete the identification of the ubiquitin family in G.elata. Finally, to identify and classify transcription factors and protein kinases in G.elata, we employed the iTAK software for comprehensive identification and classification [27].

Protein-protein interaction (PPI) network construction

Utilizing orthologous protein relationships, we aligned the protein-protein interaction (PPI) networks of A.thaliana with those of G.elata to construct the PPI network for G.elata.

Co-expression network construction

The transcriptome data were mapped to the G.elata reference genome using HISAT2 [28], and expression values were calculated with StringTie [29] to create an expression matrix. Gene correlations were then determined using the Pearson correlation coefficient, identifying highly correlated genes as co-expressed pairs. These gene correlations were ranked using the MR algorithm. The formula used for the calculation is as follows:

In the formula, “X_i"and “Y_i"denote TPM values, while “n” is the total RNA-seq sample count. “Rank” indicates the PCC order, with “AB” as gene A’s rank among genes with gene B, and “BA” as the reverse.

In co-expression networks, gene pairs exhibiting similar expression patterns are likely to share similar functions, which can be assessed using Gene Ontology (GO) annotations. The reliability of the co-expression network is enhanced when the GO similarity between co-expressed gene pairs is high. Prediction outcomes are subsequently input into a binary classifier, where the true positive rate (TPR) and false positive rate (FPR) are computed and represented as a Receiver Operating Characteristic (ROC) curve. A larger area under the curve (AUC) indicates superior predictive performance and greater stability of the co-expression network. We conducted evaluations of AUC values across various thresholds.

Using the Cfinder software [30], we identified co-expression modules, which represent groups of genes exhibiting similar expression patterns within the network. We set the parameter k = 4, indicating that each module corresponds to a complete subgraph (clique) of four genes that are all interconnected in the co-expression network. Additionally, we annotated the functions of these modules using gene set enrichment analysis with the ClusterProfiler package in R.

Identification of differentially expressed genes (DEGs)

Differential gene expression analysis was conducted using the limma package in R [31], based on transcriptomic data from public platforms. The package utilizes empirical Bayesian methods to calculate p-values and applies the Benjamini-Hochberg method to compute FDR values. Genes with an FDR < 0.05 and |log₂(fold change)| >1 were identified as differentially expressed.

Development of GelFAP v3.0

We developed the GelFAP v3.0 platform using Linux, Apache, MySQL, and PHP. A MySQL database was created to import data such as gene function annotation, PPI network, co-expression network, and gene family information. For enhanced data visualization, we built dynamic websites using HTML, PHP, JavaScript, and CSS. Platform integrated tools include gene set enrichment analysis (GSEA), JBrowse v2 for omics data visualization [32], ViroBlast for Blast analysis [33], a heatmap analysis tool, Primer design, multiple sequence alignment, DEM analysis and a Perl script for sequence extraction.

Cloning of the GeMYB4 gene

Total RNA was extracted from fresh G.elata tubers using the guanidinium isothiocyanate-phenol method (RNAiso Plus, Takara), following the manufacturer’s instructions. RNA quality and concentration were assessed using Nanodrop spectrophotometer. First-strand cDNA synthesis was performed using the PrimeScript RT reagent kit (Takara) according to the protocol provided. The GeMYB4 gene was amplified by RT-PCR using gene-specific primers designed based on the reference sequence (primer sequences are listed in Table S2). The reaction program was: 94℃ for 1 min 30 s; 30 cycles (94℃ for 20 s, 55℃ for 20 s, 72℃ for 1 min); 72℃ for 5 min; stored at 12℃. PCR products were purified by gel extraction and verified by sequencing.

Transformation of GeMYB4 and construction of transgenic materials

Following the PCR amplification and gel extraction purification of the GeMYB4 gene, the product was ligated with the pBI121 vector that had been double digested with SamI and SacI through the homologous recombination kit (pEASY^®-Basic Seamless Cloning and Assembly Kit). The ligation product was subsequently transformed into Trans1-T1 competent Escherichia coli. Positive clones were then identified on kanamycin-resistant plates and verified by colony PCR for the correct recombinant plasmid. The positive recombinant plasmid was then transformed into Agrobacterium GV3101, and single colonies were selected and grown in YEB liquid medium containing rifampicin and kanamycin until the OD₆₀₀ reached 0.6–0.8. The inflorescences of the wild-type Arabidopsis were then infected by the floral dip method. Subsequent to the infection, the plants were cultivated in accordance with standard protocols until the T0 generation seeds were harvested and sown on 1/2 MS medium containing 40 mg/L kanamycin for the purpose of resistance screening. The surviving seedlings were then transplanted and self-pollinated to obtain the T1 generation. Through a process of resistance screening and PCR verification across three consecutive generations, the presence of heterozygotes was progressively diminished, ultimately yielding homozygous lines in the T3 generation. The genomic DNA of the transgenic lines was extracted using the Fast Extraction Buffer method, and the insertion of the exogenous fragment was verified by PCR amplification. Concurrently, RNA was extracted using the RNAiso Plus method and reverse transcribed into cDNA for PCR amplification. The products were then subjected to agarose gel electrophoresis for further analysis. The transgenic lines exhibited the GeMYB4 target band, while the wild type exhibited no band, thereby confirming the integration and expression of the exogenous gene.

Determination of total flavonoid content

Transgenic GeMYB4 and wild-type Arabidopsis plants were cultivated in a light culture room at Guizhou University of Traditional Chinese Medicine. The room was kept under controlled conditions of 16 h of light and 8 h of darkness. Rutin was used as the standard control substance, and the total flavonoids content was determined by the modified aluminum chloride (AlCl₃) colorimetric method. This study examined wild-type (WT) Arabidopsis and three independent transgenic strains that all overexpress the GeMYB4 gene. 0.15 g of GeMYB4 transgenic and WT Arabidopsis aerial tissue samples were accurately weighed, crushed by liquid nitrogen, and 7.5 mL of 70% ethanol solution was added for sonication-assisted extraction at 37 °C for 30 min. 0.5 mL of the extract was filtered through a sterile syringe filter, 0.5 mL of AlCl₃ solution was added sequentially, vortex and mixed well, and then 1mL of sodium acetate-acetic acid (NaAc-HAc) buffer was added, and vortex again. After the mixture was allowed to develop color at room temperature for 15 min, the absorbance value of 415 nm wavelength was measured, and the total flavonoid content was calculated by the standard curve method.

Results

Sequence information statistics

We obtained genomic data from the National Genomics Data Center (NGDC) for the following datasets: Gel_v2.0, G.elata-2017, G.elata 2021, and G.elata.G03. The Gel_v2.0 dataset comprises 19,504 genes, 33,561 transcripts, and 33,561 proteins. In comparison, the G.elata-2017 dataset contains 19,637 genes, 18,969 transcripts, and 18,969 proteins. The G.elata 2021 dataset includes 19,577 genes, 36,918 transcripts, and 36,918 proteins. Lastly, the G.elata.G03 dataset consists of 17,895 genes, 26,453 transcripts, and 26,453 proteins, as illustrated in Fig. 1A. 

Fig. 1.

Overview of the Functional Genomics Database. A Gene function annotation information. From top to bottom, the full names are coding sequences (CDS), transcript variants (Transcript), genes, and translated EMBL Nucleotide Sequence Database (trEMBL), Non-redundant Protein Sequence Database (NR), Arabidopsis Information Resource (TAIR10), Pfam Protein Family Database (Pfam), Swiss-Prot, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG). B Gene family classification. From top to bottom, the full names are transcription factor and regulator (TF&TR), protein kinase (PK), ubiquitinating enzymes (ubiquitin), transporter family (TP), carbohydrate-active enzyme (CAZy), cytochrome P450 (CYP450)

Gene family classification

First, we used iTAK software to analyze protein kinases (PKs) and transcription factors (TFs) in four versions of G.elata, and predicted the potential PK numbers of 573, 584, 540, and 572 in the Gel_v2.0, G.elata-2017, G.ealta 2021, and G. elata.G03 four versions, and 1,277, 1,297, 1,219, and 1,253 potential TFs, respectively. Second, we predicted the ubiquitin-proteasome-encoding genes in G.elata based on the ubiquitin-proteasome hidden Markov model (HMM) downloaded from the iUUCD v2.0 database. Finally, we compared all genes to the CAZy database, and the results showed that the number of CAZy in the four versions of G.elata was 423, 609, 403, and 439, respectively (Fig. 1B).

Functional annotation

We compared the protein sequences of four G.elata versions with the NR, TAIR, Swissprot, and trEMBL databases. Using the NR database, we annotated 17,303, 15,285, 15,395, and 14,941 genes, respectively, and with the TAIR database, we annotated 14,301, 13,864, 13,044, and 13,453 genes, respectively. The four G.elata versions were aligned with the Swissprot database, annotating 12,882, 11,993, 11,496, and 11,669 genes, and with the trEMBL database, annotating 17,305, 15,314, 15,403, and 14,951 genes, respectively. We annotated genes in four G.elata versions using the GhostKOALA tool from the KEGG database, covering 2,785, 3,188, 2,957, and 8,163 genes respectively. Additionally, we used PfamScan to functionally annotate protein domains for 13,692, 12,267, 12,255, and 12,097 genes. We used InterProScan for GO annotation, resulting in 13,692, 12,267, 12,255, and 12,097 genes annotated across four G.elata versions (Fig. 1A).

Protein-protein interaction (PPI) networks

Based on the constructed protein-protein interaction (PPI) networks in Arabidopsis thaliana, we mapped these networks to G.elata. Consequently, we developed a PPI network for G.elata comprising 6,920 nodes and 36,467 edges in the Gel_v2.0 version, 5,726 nodes and 28,816 edges in the G.elata-2017 version, and 7,038 nodes and 39,881 edges in the G.elata 2021 version. Additionally, the G.elata.G03 version contains 6,591 nodes and 40,937 edges, as illustrated in Fig. 2D.

Fig. 2.

Statistics of PCC distribution, co-expression network, and PPI network data for the platform. A The relationship between the Pearson correlation coefficient (PCC) and the number of edges in the co-expression network. B Statistical analysis of nodes and edges in positive co-expression network. C Statistical analysis of nodes and edges in negative co-expression network. D Statistical analysis of nodes and edges in protein-protein interaction (PPI) network

Co-expression network

We mapped RNA-seq reads from the SRA database to the G.elata genome, selected transcriptomes with mapping rates over 70% (Table S1), and calculated gene expression levels for all samples. We analyzed Pearson correlation coefficients (PCCs) from expression profiles to identify co-expressed gene pairs(Fig. 2A). Consequently, the MR method was employed to further elucidate the closely interconnected gene pairs within the network.

To ensure the reliability of the network we constructed, we compared the AUC values at different PCC thresholds (0.6, 0.7, 0.8, 0.9) and found that, regardless of the genome version, the AUC value was highest when the PCC exceeded 0.8 (Fig. S1). The specific method can be found in the supplementary information. Similarly, we analyzed the AUC values for MR20, MR30, MR40, and MR50 and found that, regardless of the genome version, the AUC value was highest when MR < 30 (Fig. S2). Based on the above results, we selected PCC > 0.8 and MR < 30 to construct the positive co-expression network. Finally, we obtained a positive co-expression network with 12,545 nodes and 94,164 edges in Gel_v2.0, a positive co-expression network with 12,604 nodes and 90,854 edges in G.elata-2017, and a positive co-expression network with 12,252 nodes and 93,566 edges in G.ealta 2021. G.elata.G03 has a positive co-expression network of 11,710 nodes and 87,181 edges (Fig. 2B). At the same time, Gel_v2.0 also obtained a negative co-expression network with 7927 nodes and 44,518 edges, G.elata-2017 with 8972 nodes and 50,463 edges, and G.ealta 2021 obtained a negative co-expression network with 7476 nodes and 41,675 edges. G.elata.G03 has a negative co-expression network of 7892 nodes and 43,454 edges (Fig. 2C).

Functional modules identification and annotation

Using the CFinder software, we identified functional modules from the gene co-expression networks for four different G.elata genomes. Specifically, 290, 306, 309, and 240 modules were identified for G.elata-2017, Gel_v2.0, G.elata 2021, and G.elata.G03, respectively. Gene set enrichment analysis was conducted using the ClusterProfiler package in R to functionally annotate each module. Based on Gene Ontology (GO) enrichment, the annotation coverage—defined as the proportion of modules with at least one significantly enriched GO term (adjusted p-value < 0.05)—was 59.7% (173/290), 61.8% (189/306), 54.7% (169/309), and 60.0% (144/240), respectively. For KEGG Orthology (KO) enrichment, the corresponding coverage rates were 59.3% (172/290), 49.7% (152/306), 49.2% (152/309), and 61.7% (148/240), respectively. These results indicate that a substantial proportion of the co-expression modules are functionally enriched, supporting the biological relevance and robustness of the co-expression network construction across different genome assemblies.

Construction of GelFAP v3.0

We have developed an updated version of the G.elata gene function analysis platform, GelFAP v3.0, through comprehensive gene family classification, functional annotation, protein-protein interaction (PPI) network analysis, and co-expression network analysis. In GelFAP v3.0, we have made substantial improvements in both data scale and quality. Specifically, the number of supported genomes has increased from a single genome in previous versions to four genomes, and the number of transcriptome samples has expanded from 39 to 51 in GelFAP v1.0 and v2.0, respectively, to 95 transcriptomes in the current version. Regarding genome quality, the newly integrated G.elata 2021 genome version demonstrates significantly higher N50 values compared to the genomes used in previous versions. Furthermore, the BUSCO completeness score of the G.elata.G03 genome is also superior, indicating improved assembly integrity and annotation quality (Tables S3, S4). In terms of functional enhancements, the current version offers a more comprehensive set of analysis tools, including data download, module search, expression profile visualization, tissue-specific pathway information, and a more user-friendly and visually appealing interface. This platform comprises eight distinct sections: Home, Genomes, Network, Search, Tools, Pathway, Download, and Manual. Each section offers researchers valuable insights and resources for in-depth gene function analysis. The genome provides detailed information on the four versions of the G.elata genome, covering an overview, description, assembly, and family classification for each. The network features a co-expression network and a PPI network. Users can enter a keyword and accession number for a gene, transcript, or protein on the search page to find specific genes. The tools section comprises nine branches: Blast, Enrichment, Extract sequence, JBrowse v2, Multiple sequence alignment, Primer design, DEG analysis, DEM analysis and Heatmap Analysis, while the Pathways section features KEGG-annotated pathway genes as well as special pathways. Users can download data on the genome, gene family, functional annotation, co-expression network, PPI network, and sequences for the four G.elata versions. A user manual is available to help new users quickly understand the site’s features. Figure 3 shows the complete architecture of the built GelFAP v3.0 platform (Fig. 3).

Fig. 3.

Organizational chart of GelFAP v3.0

Network viewer of DEGs

We detected DEGs across seven transcriptome datasets, yielding 82 DEG clusters (Table S5). For better display gene network and expression, we build a network viewer to merge gene co-expression and PPI networks with DEG. The integrated network visualization depicts up-regulated (red) and down-regulated (blue) DEGs in network display pages.

Differential expression analysis of metabolites

We analyzed and organized differentially expressed metabolites under various conditions, resulting in information on 10 sets of such metabolites (Table S6). We developed a network viewer that provides users with volcano plots of differentially expressed metabolites and detailed information for each on the webpage, enhancing the presentation of G.elata metabolomics data. In the integrated network visualization, upregulated metabolites are highlighted in red, while downregulated metabolites are shown in blue.

Case study 1: functional analysis of MYB4 gene in Gastrodia elata

MYB4 gene is an important transcription factor in plants, belonging to the R2R3-MYB family, and is involved in the regulation of various biological processes such as secondary metabolism and stress responses. For instance, in Arabidopsis, MYB4 inhibits the accumulation of flavonoids by directly binding to the promoters of the 4CL and FLS genes [34, 35]. Moreover, MYB4 in plants also plays roles in enhancing tolerance to cadmium stress [36], participating in DNA damage repair [34], responding to drought stress [37], cold stress [38], salt stress [39] and osmotic stress [40]. To demonstrate the platform’s role in assisting users with gene function research, we used the MYB4 gene as a case study to search GelFAP v3.0. The search revealed a gene with the identifier GWHGBDNU013767 as the closest match. The platform’s information indicates that this gene is annotated as myb-related protein Zm38 and MYB domain protein 4 (Fig. 4A). The MYB4 gene is located on the positive strand of chromosome 4, spanning from position 66,334,482 to 66,335,611 base pairs (Fig. 4B). The Gene detail interface also provides a link to the co-expression network (Fig. 4C). Protein domain analysis indicates that this gene contains two MYB domains (Fig. 4D). Gene family classification information reveals that the gene belongs to the MYB transcription factor family (Fig. 4E). Additionally, the expression profile provides the expression levels of MYB4 across different transcriptomes (Fig. 4F).

Fig. 4.

Gene details page for the GeMYB4 gene. A Gene function annotation. B Location and transcription sequence. C GeMYB4 network. D Protein structure and sequence. E Gene family classification. F Expression levels in different samples

Previously reported that the total content of p-hydroxybenzyl alcohol in tuber was higher than that in flower [12]. The expression data provided by the platform indicated that flower exhibited significantly higher expression levels compared to tuber (Fig. 4F and 5A). Through co-expression network analysis, we found that only one gene had a co-expression relationship with GeMYB4. This gene is annotated as a Transmembrane amino acid transporter family protein and has a co-expression relationship with GWHGBDNU001354 and GWHGBDNU001355. Annotation information indicates that both of two genes are Chalcone and stilbene synthase family proteins, which may be related to the biosynthesis of flavonoids. We further analyzed the top 300 genes in terms of correlation (Table S7). One of them is the MYBAS1-like myb-related protein (GWHGBDNU001715), and its TAIR annotation information also supports its participation in the flavonoid biosynthesis process. Diacylglycerol O-acyltrantferase 1 (GWHGBDNU016109) and cellulose synthase-like D3 (GWHGBDNU011829) play regulatory roles in the cold response process. GWHGBDNU016109 and GWHGBDNU005362 play regulatory roles in responding to salt stress. These analysis results are also supported by several prior knowledge.

Fig. 5.

Functional analysis of the GeMYB4 gene. A Expression of GeMYB4 in tubers and flowers. B GWHGBDNU013767 is displayed with a positve co-expression network of GWHGBDNU006344. C EAR-motif of GeMYB4 after multi-sequence alignment with AtMYB3, AtMYB4, AtMYB7, and AtMYB32. D Schematic diagram of GeMYB4 overexpression vector. E Identification of positive clones of GeMYB4 gene. F Total flavonoid content of GeMYB4 transgenic plants and WT wild-type. The symbols ** and *** indicate that the total flavonoid content shows highly significant (p < 0.01) and extremely significant (p < 0.001) differences between GeMYB4 and WT, respectively

The EAR motif represents a conserved functional domain in plant transcription factors, with core sequences typically defined as “LxLxL” or “DLNxxP” (where x denotes variable amino acid residues). Its primary function acts as a transcriptional repression domain: by recruiting inhibitory complexes like histone deacetylases (HDAC), it alters chromatin conformation and suppresses the transcriptional activity of target genes [41]. Through multi-sequence alignment of the GeMYB4 protein sequence with Arabidopsis proteins containing the EAR motif, we identified an EAR motif at the C-terminus of GeMYB4 (Fig. 5C). Thus, GeMYB4 may exert inhibitory effects during flavonoid synthesis. To further validate the inhibitory role of GeMYB4, we constructed the overexpression vector 35S: GeMYB4 (Fig. 5D). Positive plants were subsequently generated via overexpression in Arabidopsis, and RT-PCR detection confirmed their positivity (Fig. 5E). Final analysis results demonstrate that GeMYB4 likely inhibits the biosynthesis process of flavonoids (Fig. 5F).

Case study 2: a functional analysis of the Gastrodia elata PAL1 gene

Phenylalanine ammonia-lyase (PAL) represents the initial enzyme in the phenylpropanoid pathway, which has the capacity to generate precursors for a range of significant secondary metabolites, including various phenolic compounds [42]. In plants, the PAL gene is encoded by a multi-gene family. For example, in the Arabidopsis genome there are four PAL homologous genes [43]. Research has demonstrated that PAL1 plays a pivotal role in the synthesis of phenolic compounds, thereby enhancing the accumulation of polyphenolic compounds in rapeseed. This process is facilitated by the promotion of the phenylpropanoid metabolic pathway under conditions of drought stress [44]. Studies have shown that G.elata. The content of p-hydroxybenzyl alcohol in G.elata f.glauca (GB, black stem) was higher than that of G.elata f.elata (GR, red stem) [45]. Through the GelFAP v3.0 platform, we searched for a gene with the identifier GWHGBDNU010866 (GePAL1), which best matched PAL1 according to annotations. The results of the platform demonstrated that the expression level of GePAL1 in GB was significantly higher than that of GR (Fig. 6A), and we also performed co-expression analysis of the GePAL1 gene, which showed that it was positively co-expressed with 5 genes and negatively co-expressed with 5 genes (Fig. 6B). The conditional co-expression network analysis between GB and GR revealed that 5 genes were positively co-expressed with GePAL1, and one gene, GWHGBDNU010861, was significantly up-regulated. These results suggest a potential synergy between GePAL1 genes and the GePAL1 gene in the synthesis of phenolic compounds (Fig. 6C). Then, we analyzed the positive co-expression genes of GWHGBDNU010866, and the results showed that the GWHGBDNU010866 gene and its positive co-expressed genes were highly expressed in GB, while their expression levels were significantly reduced in GR (Fig. 6D). Therefore, our analysis results suggest that the GWHGBDNU010866 gene plays an important regulatory role in the synthesis of hydroxybenzyl alcohol content, suggesting that the PAL1 gene may be involved in the synthesis of p-hydroxybenzyl alcohol.

Fig. 6.

Functional analysis of the GWHGBDNU010866 gene. A Expression of GWHGBDNU010866 in Gastrodia elata f.glauca and Gastrodia elata f.elata. B Positive co-expression network and negative co-expression network of the GWHGBDNU010866. Orange links represent positive co-expression relationships, green links represent negative co-expression relationships. C Co-expression network of GWHGBDNU010866 in Gastrodia elata f.glauca vs. Gastrodia elata f.elata. Red nodes represent up-regulated genes, green represent no significant change genes, and orange links represent positive co-expression relationships. D Heatmap of GWHGBDNU010866 and its positive co-expressed gene

Case study 3: functional analysis of module0023 implicated in fungal response

Within the module search interface, we found a functional module designated module0023, which may play a role in fungal defense responses according GO enrichment analysis result. The module under consideration contains five genes: drought-induced 21, ABC transporter A family protein, translocase inner membrane subunit 17 − 2, Alba DNA/RNA-binding protein, and one gene with unknown annotation. It is noteworthy that two of these genes exhibit significant upregulation in mature tubers infected by fungi compared to healthy, mature tubers. This expression pattern suggests the possibility of synergistic functionality among the genes contained within module0023, which may be implicated in the plant’s innate defense mechanism against fungal infections (Fig. S3).

Discussion

As a common traditional Chinese medicine, G.elata has been included in the list of “medicine and food”, and G.elata is also used in some health products. In recent years, six versions of the G.elata genome have been released, and the transcriptome data of G.elata have been continuously updated and accumulated. Therefore, we have integrated four versions of G.elata genome data, transcriptome data, metabolomics data, processed annotation data and analysis tools, and updated the G.elata gene function analysis platform (GelFAP v3.0). At present, many species platforms have been released to collect and analyze gene function information of related species, such as MangroveDB [46], SpudDB [47], Gymnotoa-DB [48], Pear Genomics Database [49], etc., but these databases are based on crops, plants, fruits, etc. The GelFAP v3.0 we built is based on medicinal plants, which can provide reference for the subsequent construction of gene function platforms for other medicinal plants. Compared with the previous two versions of G.elata gene function analysis platform, the current version can provide updated genomic data, more complete transcriptome resources and more analysis tools, which can provide certain data resources and convenient tools for researchers who take G.elata as the research object. For example, in the GelFAP v3.0 platform, we have added an online primer design feature, making it easier for users to go from sequence browsing to primer design, thus achieving a full-cycle solution for molecular experiment needs.

Plant MYB transcription factors constitute a large family that plays essential roles in numerous physiological and biochemical processes. Flavonoids are a class of sugar-based chemical compounds characterized by a flavonoid structure, which are widely distributed in plants and exhibit a variety of biological activities and pharmacological effects. These compounds play essential roles in plant growth, development, defense mechanisms, and other physiological processes [50]. Transcriptional control is a central regulatory mechanism in the modulation of flavonoid biosynthesis [51]. Multiple studies have established that MYB transcription factors contribute to the regulation of flavonoid metabolism. Ectopic expression of BsMYB36 and BsMYB51 led to a marked elevation in flavonoid content in Bletilla striata [52]. EbMYBP1, a transcription factor isolated from Erigeron breviscapus, has been shown to directly activate flavonoid biosynthetic genes, thereby playing a regulatory role in flavonoid production [53]. Functional characterization reveals that NtMYB4a promotes anthocyanin accumulation and participates in abiotic stress responses in tobacco plants [54]. Studies have found that AtMYB4 can suppress flavonoid accumulation by repressing the expression of Arogenate Dehydratase 6 (ADT6)-encoding gene [35]. In this study, We found a GeMYB4 gene through the platform, and based on the information provided by the platform, we speculate that this gene may also play a role in inhibiting flavonoid biosynthesis. These results indicated that GeMYB4 may be a negative regulator of flavonoid biosynthesis pathway, which can provide a certain reference for subsequent research on flavonoid biosynthesis pathway in G.elata. In the future, we will be able to screen the regulatory factors that interact with GeMYB4 through yeast two-hybrid experiments, which is conducive to removing the inhibition of flavonoid biosynthesis in G.elata by GeMYB4.

While our GelFAP v3.0 provides users with valuable information and tools, we also recognize that there are still areas for improvement in the platform. We have integrated the updated genome and transcriptome data of G.elata, but with the rapid development of high-throughput sequencing, Comprehensive data on G.elata will be released progressively. Upcoming releases of G.elata epigenomic and proteomic datasets will be systematically integrated into our platform.Therefore, we plan to continue to integrate the effective and timely collection and processing of multi-omics data on G.elata in the future, update the relevant information, and upgrade the functions of the tools in the platform, so as to provide more references for future research on G.elata. And with the rapid development of artificial intelligence technology, it has had a profound impact on the construction, management and application of our biological database. In the future, our platform will also integrate some intelligent analysis tools and applications in the future, such as users can change from manual information retrieval to obtain relevant information through natural language dialogue. If users or researchers have the latest data they wish to display on the platform, they can contact us through the developer email provided. We would greatly appreciate this contribution.

In conclusion, GelFAP v3.0 is a dedicated platform for investigating functional genes and associated biological processes in G.elata, offering analytical capabilities including co-expression network construction and gene enrichment analysis. The platform will be continuously updated with with the latest multi-omics data to support users’ research on G.elata. The platform is available for free at http://www.gzybioinformatics.cn/Gelv3.

Authors’ contributions

The development of the database was completed by BP, JY, and QP. BP, JZ, and PZ participated in the conduct of the relevant experiments. JY and JL provided research ideas and financial support. The first draft of the manuscript was written by BP, and revised by JY, JZ, JY, and HY. The final version of the manuscript has been reviewed and approved by all authors.

Funding

This work was supported by the National Natural Science Foundation of China (No. 32160139), the Guizhou University of Traditional Chinese Medicine Graduate Education Innovation Program (YCXKYS2023029) and the National and Provincial Scientific and Technological Innovation Talent Team of the Guizhou University of Traditional Chinese Medicine (GZYTDHZ[2022]003, GZYTDHZ[2024]002).

Data availability

Genomic data and genome annotation were obtained from the National Genomics Data Center (NGDC) (accession number: GWHBDNU00000000, GWHAAEX00000000, GWHAOSR00000000, GWHBHOU00000000). Transcriptomic data are list in supplementary table S1.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.