PlantMetaDB

1.1 About PlantMetaDB

PlantMetaDB (Plant Metabolic Database) is a comprehensive database specifically designed for plant natural product metabolism. It integrates four core data modules: enzymes, reactions, compounds, and species. The database aims to provide high-quality, structured data support for plant metabolism research, biosynthetic pathway reconstruction, and natural product discovery. Users can access the core functions via: https://www.np-design.cn/plantmetadb/

Plant natural products (such as alkaloids, terpenoids, and flavonoids) are important sources for pharmaceuticals, nutraceuticals, and industrial materials. However, many high-value plant natural products have extremely low yields in their native hosts and complex molecular structures, making large-scale chemical synthesis difficult. PlantMetaDB not only systematically integrates data from public databases such as KEGG, MetaCyc, RHEA, UniProt, and LOTUS, but more importantly, we have performed deep data processing and functional extensions:

1.2 Database Statistics

PlantMetaDB contains rich plant metabolic data:

1.3 Key Features

• Multi-Dimensional Data Integration: Links enzyme, reaction, compound, and species data through standardized identifiers.
• AI-Assisted Annotation: EC number prediction using HDMLF deep learning model, greatly expanding enzyme functional annotation coverage.
• Taxonomic Traceability: Based on NCBI taxonomy, supports hierarchical browsing from family to genus to species.
• Chemical Classification Integration: Systematic compound classification using NPClassifier.
• Sequence Analysis Tools: Built-in BLAST sequence alignment and EC number prediction functionality.

2.1 Homepage

The PlantMetaDB homepage displays core database statistics and main function entries. The top navigation bar contains the following main module links:

• Home: Database introduction and statistics overview.
• Reaction: Metabolic reaction query and browsing.
• Enzyme: Enzyme data query, sequence alignment, and EC number prediction.
• Compound: Plant natural product compound query.
• Organism: Plant species taxonomy browsing.
• Data: Website data update records.
• Help: User manual and contact information.

2.2 Database Identifier System

PlantMetaDB uses a unified identifier system to manage data entries:

• PME (PlantMetaDB Enzyme): Enzyme entry identifier, e.g., PME000001
• PMR (PlantMetaDB Reaction): Reaction entry identifier, e.g., PMR000001
• PMC (PlantMetaDB Compound): Compound entry identifier, e.g., PMC000001
• PMO (PlantMetaDB Organism): Species entry identifier, e.g., PMO000001

3.1 Enzyme Data Search

The Enzyme module provides multiple search methods to help users quickly find target enzyme information.

3.1.1 Basic Search

Users can search by the following fields:

• PME ID: PlantMetaDB enzyme identifier, e.g., PME000001
• UniProt ID: UniProt database identifier, e.g., P12345
• Species Name: Scientific name of the plant species, e.g., Arabidopsis thaliana
• EC Number: Enzyme classification number, e.g., 2.1.1.104

3.1.2 EC Classification Selection

EC numbers (Enzyme Commission numbers) have a hierarchical structure. Users can select level by level via dropdown menus:

• Level 1 (Class): EC 1 (Oxidoreductases), EC 2 (Transferases), EC 3 (Hydrolases), etc.
• Level 2 (Subclass): e.g., EC 2.1 (Transferring one-carbon groups)
• Level 3 (Sub-subclass): e.g., EC 2.1.1 (Methyltransferases)
• Level 4 (Specific enzyme): e.g., EC 2.1.1.104

3.1.3 Review Status Filter

Users can filter by the review status of enzyme entries:

• All: Display all enzyme entries.
• Reviewed: Only display enzymes marked as reviewed in UniProt (Swiss-Prot). These entries have been manually curated by experts and experimentally validated with high data reliability.
• Unreviewed: Display enzymes marked as unreviewed in UniProt (TrEMBL). These entries are computationally annotated or model-predicted and have not been manually validated.

3.2 Sequence BLAST Alignment

The Enzyme module provides sequence alignment functionality, allowing users to input protein sequences for similarity comparison with enzymes in the database.

3.2.1 How to Use

1. Click the "Sequence BLAST" tab to switch to the sequence alignment interface.
2. Enter the protein sequence in the text box (FASTA format or plain sequence are both accepted).
3. Click the BLAST Search button to start alignment.
4. Wait for alignment to complete and view the similarity results list.

3.2.2 Input Format

Supported sequence input formats:

• FASTA format: Header line starting with ">" followed by the sequence.
• Plain sequence: Directly paste amino acid sequence (without header line).

3.3 EC Number Prediction

Based on the ECRECer deep learning model, users can predict EC number classification for unknown sequences.

3.3.1 How to Use

1. Click the EC Prediction tab.
2. Enter the protein sequence to be predicted in the text box.
3. Click the Predict EC button.
4. View prediction results, including EC number and confidence score.

3.3.2 Result Interpretation

Prediction results contain the following information:

• EC Number: Predicted enzyme classification number; click to navigate to ExPASy for details.
• Confidence: Prediction reliability: ≥70% is high confidence, 30–70% is medium, <30% is low confidence.
• Enzyme Class Description: Enzyme function description for the corresponding EC number.
• View Reactions: Click to view known reactions for this EC number.

3.4 Extended Search Result Features

3.4.1 Sorting by Confidence

A sort button (⇅) is provided to the right of the "Confidence" column header in the result list. Clicking it cycles through the following three states:

• Default order (original database sort)
• Confidence from high to low (↓)
• Confidence from low to high (↑)

3.4.2 Exporting Results

The toolbar above the search result list provides two data export formats:

• Export CSV: Export as a comma-separated text file, suitable for programmatic processing.
• Export Excel: Export as Excel XML format (.xls), suitable for opening directly in Office.

The exported content reflects the result data under the current search conditions.

4.1 Reaction Data Search

The Reaction module allows users to query and browse plant metabolic reaction data.

4.1.1 Search Options

Users can search by the following fields:

• PMR ID: PlantMetaDB reaction identifier, e.g., PMR000001
• RHEA ID: RHEA database reaction identifier
• KEGG ID: KEGG database reaction identifier, e.g., R00001
• MetaCyc ID: MetaCyc database reaction identifier
• EC Number: EC classification number of the enzyme catalyzing the reaction

4.1.2 EC Number Hierarchical Search

The Reaction module supports four-level precise input of EC numbers. Users can specify each level of the EC number hierarchically (e.g., "2.1.1.-" represents all methyltransferase reactions). The wildcard "-" can be used to match all subclasses at that level.

4.2 Reaction Details Page

Click on a reaction entry to view detailed information, including:

• Reaction Equation: Chemical structure diagrams of substrates and products.
• Main Components: Main substrates and products involved in the reaction.
• Cofactors: Coenzymes/cofactors required for the reaction (e.g., ATP, NADH).
• Catalytic Enzymes: List of enzymes catalyzing the reaction.
• Species Source: Plant species information involved in the reaction.
• External Links: Links to RHEA, KEGG, MetaCyc, and ExPASy external databases.

4.3 Chemical Structure Drawing Search

The Reaction module provides a chemical structure visual search function based on the Ketcher editor. This area is collapsed by default; click the "Chemical Structure Search" title bar to expand it. The interface contains two independent molecular structure drawing panels for specifying the Reactant and Product structures. Each panel supports the following operations:

• Draw Structure: Draw chemical structures directly in the embedded Ketcher editor.
• Input SMILES: Paste the SMILES string in the input box and click "Load SMILES" to load it into the editor.
• Get SMILES: Click "Get SMILES" to export the current drawn structure as a SMILES string.
• Clear: Click "Clear" to clear the current drawing.

Each panel also provides a search mode toggle button to switch between the following two modes:

• Exact Search: Returns only reactions involving compounds that exactly match the drawn structure.
• Substructure Search: Returns reactions involving compounds that contain the drawn structure as a substructure.

After completing the drawing, click the "Search by Chemical Structure" button to execute the search. Reactant and product conditions can be specified individually or simultaneously.

5.1 Compound Data Search

The Compound module provides search and browsing functionality for plant natural products.

5.1.1 Search Options

• Compound Name: Enter the English name or common name of the compound for fuzzy matching.
• Species Name: Filter compound sources by plant species.
• Chemical Classification: Filter based on NPClassifier classification system, including:
  • Terpenoids: 49.4% of total compounds
  • Phenylpropanoids: 23.7%
  • Alkaloids: 11.7%
  • Polyketides: 5.2%
  • Fatty acids: 3.9%
• Molecular Weight Range: Specify minimum and maximum molecular weight for filtering.

5.1.2 Search Result Display

Search results are presented in tabular form. Each record contains the following fields: PMC ID, molecular structure image (Mol Structure), compound name (Name), SMILES string, molecular formula / molecular weight (Formula / MW), chemical classification (Category), species origin (Species), and literature DOI.

5.2 Compound Details Page

The compound details page displays the following information:

• 2D Structure Diagram: Two-dimensional molecular structure of the compound.
• Molecular Formula and Weight: Molecular composition information of the compound.
• NPClassifier Classification: Hierarchical classification (Pathway → Superclass → Class).
• Species Source: Plant source of the compound and complete taxonomic information.
• Related Reactions: Metabolic reactions involving this compound.
• Literature Reference: Original literature DOI link.

6.1 Taxonomic Hierarchical Browsing

The Organism module uses the NCBI taxonomy system, supporting four-level hierarchical browsing:

• Plant Type: Angiosperms (91.6%), Gymnosperms, Pteridophytes, Bryophytes
• Family: e.g., Asteraceae, Fabaceae, etc.
• Genus: e.g., Arabidopsis
• Species: e.g., Arabidopsis thaliana

6.2 How to Use

1. Select plant type (e.g., Angiosperms).
2. Select target family from the displayed family list.
3. Continue selecting genus and species.
4. After selection, view related data for that taxonomic unit:
   • Related Reactions: Metabolic reactions recorded in that species/taxonomic group.
   • Related Enzymes: Enzymes sourced from that species/taxonomic group.
   • Related Compounds: Natural products detected in that species/taxonomic group.

6.3 Species Metabolic Data Characteristics

Different species in the database have varying levels of metabolic data coverage:

Term	Description
EC Number	Enzyme classification system established by the International Union of Biochemistry
Plant Natural Products	Secondary metabolites produced by plants
Terpenoids	A large class of natural products composed of isoprene units
Phenylpropanoids	Compounds derived from phenylalanine
Alkaloids	Nitrogen-containing plant secondary metabolites
Metabolic Pathway	A series of consecutive biochemical reactions
Oxidoreductases	Enzymes catalyzing oxidation-reduction reactions (EC 1)
Transferases	Enzymes catalyzing group transfer reactions (EC 2)
Hydrolases	Enzymes catalyzing hydrolysis reactions (EC 3)
Lyases	Enzymes catalyzing non-hydrolytic cleavage reactions (EC 4)
Isomerases	Enzymes catalyzing isomerization reactions (EC 5)
Ligases	Enzymes catalyzing ligation reactions (EC 6)
Translocases	Enzymes catalyzing the translocation of ions or molecules across membranes (EC 7)
Swiss-Prot	Manually curated protein dataset in UniProt (corresponding to Reviewed status)
TrEMBL	Automatically annotated protein dataset in UniProt (corresponding to Unreviewed status)