SEARCH EDCs

Users can opt for Search option in the navigation bar to obtain information on EDCs based on their chemical name, chemical identifiers such as PubChem, CAS, DSSTox or DEDuCT, or physicochemical properties. Further, one can also fetch the list of top 10 structurally similar EDCs to the query compound using chemical similarity search option.

The results of the query search will display a table containing the chemical name and chemical identifiers. Users can click on the name of an EDC from the resulting table to view the separate page containing the diverse information such as chemical structure, detailed experimental evidence, physicochemical properties, molecular descriptors, predicted ADMET properties, chemical-gene interaction, chemical-phenotype interaction, chemical-disease association, active ToxCast endpoints, associated adverse outcome pathways (AOPs), their presence in consumer products, chemical lists, and human biospecimens. The result page for a particular EDC also provides an option for the users to download its 2D and 3D structure, resulting table for predicted ADMET properties and molecular descriptors.

Simple search

Using simple search, users can retrieve EDCs in our database using the chemical name or their corresponding chemical identifiers (PubChem or CAS or DSSTox or DEDuCT).

Physicochemical filter

Using physicochemical filter, users can retrieve EDCs based on physicochemical properties such as molecular weight, logP, topological polar surface area (TPSA), hydrogen bond acceptors (HBA), hydrogen bond donors (HBD), heavy atoms, heteroatoms, and rotatable bonds.

Chemical similarity filter

The chemical similarity filter can be used to obtain top 10 EDCs that are structurally similar to the query compound in our database evaluated based on Tanimoto coefficient.

BROWSE EDCs

Users can retrieve EDCs based on the following types of classification using Browse option in the navigation bar.

• Based on type of supporting evidence for endocrine disruption
• Based on environmental source
• Based on chemical classification
• Based on systems-level perturbations and endocrine-mediated endpoints

1. BASED ON TYPE OF SUPPORTING EVIDENCE IN PUBLISHED EXPERIMENTS

Users can browse EDCs based on the type of supporting evidence for endocrine disruption from published experiments in humans or rodents. Users can also easily navigate back to the page listing all the chemicals in a group by clicking on the group information within the chemical information page.

2. BASED ON ENVIRONMENTAL SOURCE

The browse EDCs option helps the users to fetch the list of EDCs classified under any of the 7 broad categories or 48 sub-categories.

7 BROAD CATEGORIES

In this section, the users can retrieve the list of EDCs in each of the 7 broad categories. The results page provides the list of EDCs and the number of EDCs classified under each broad category. Moreover, the users can also click on the corresponding information within chemical pages, to navigate to the respective broad category pages.

48 SUB-CATEGORIES

In this section, the users can retrieve the list of EDCs in each of the 48 sub-categories. The results page provides the list of EDCs and the number of EDCs classified under each sub-category. Moreover, the users can easily navigate to the chemical list page by selecting appropriate information within the chemical information page.

3. BASED ON CHEMICAL CLASSIFICATION

The browse option based on chemical classification provides the non-overlapping two-level hierarchical classification of EDCs categorized using the structure and composition of EDCs. The tree map facilitates the users to filter out the set of EDCs based on chemical kingdom or chemical super class. The results page provides the list of EDCs and the number of EDCs present in the particular category. Moreover, the users can easily navigate back to the table information page, by clicking on the appropriate section in the chemical information page.

4. BASED ON SYSTEMS-LEVEL PERTURBATIONS AND ENDOCRINE-MEDIATED ENDPOINTS

Users can browse EDCs based on endocrine-mediated endpoints specific to 7 systems-level perturbations.

Chemical page

Associated AOPs

This page shows the various Adverse Outcome Pathways (AOPs) from AOP-Wiki (version: 2025-04-01), associated with the chemical, providing insights into potential mechanisms of toxicities. The associated AOPs are categorized based on the Level of Relevance which is defined as follows:

1. Level 1: The stressor is associated with at least one KE within an AOP, where the KE is neither MIE nor AO within that AOP
2. Level 2: The stressor is associated with at least one AO within an AOP, but not associated with any MIE within that AOP
3. Level 3: The stressor is associated with at least one MIE within an AOP, but not associated with any AO within that AOP
4. Level 4: The stressor is associated with at least one MIE and one AO within an AOP
5. Level 5: The stressor is associated with at least one MIE and one AO within an AOP and there exists a directed path between the associated MIE and AO

Here, MIE stands for Molecular Initiating Event, KE stands for Key Event, and AO stands for Adverse Outcome.

For each associated AOP, the page provides information on the corresponding identifier, title, AO classification, OECD status, coverage score, and the KEs that are mapped to the toxicological endpoints associated with the chemical.

Here, the coverage score is defined as the fraction of KEs within the AOP, that are mapped to the chemical-associated toxicological endpoints.

The 'Taxonomic applicability' column provides data on the organisms for which the AOP is applicable.

DEDuCT-KG page

Interactive DEDuCT Knowledge Graph

The DEDuCT-KG page allows the users to interact with the DEDuCT knowledge graph that is hosted in the neo4j database. This interactive user interface (UI) allows users to query individual nodes, or find the shortest paths present between any two nodes.
1. Searching & Filtering

• Node Search: Using the dropdown on the left, users can filter nodes by 7 different node types. As the user types in the search bar, the system will provide suggestions for specific identifiers as dropdown, which can then be selected by clicking on the appropriate search term. The following are the types of nodes in the knowledge graph.
  • Chemical: potential EDCs present in DEDuCT
  • DEDuCT Endpoint: endocrine-mediated endpoints curated in DEDuCT
  • Gene: chemical-associated genes curated from CTD and ToxCast
  • Key Event: measurable biological events in an adverse outcome pathway (AOP) curated from AOP-Wiki
  • AOP: Adverse Outcome Pathways (AOPs) curated from AOP-Wiki
  • Phenotype: chemical-associated phenotypes curated from CTD
  • Disease: chemical-associated diseases curated from CTD
  The initial view displays only 50 nodes related to the selected node. Users can use the Neighbor Slider to adjust the number of nodes displayed as desired.
• Shortest Path Search: Users can turn on the End Node switch to select a second node. The network then shows the shortest paths (up to four edges long) between the two selected nodes. In this mode, the neighbor node type filter is disabled.
• Node Filter: Users can filter the neighboring nodes using the dropdown titled 'Select Neighbor Node Type'. Users can select one or more types from the list. The selected filters appear as individual cards within the search field. Users can deselect a specific filter by clicking the cross on its card. All active filters can be removed at once by clicking the main cross in the search box. This filter is available for single-node searches only.
• Edge Filtering: Using the top dropdown titled 'Select Relation', users can select specific relationship types (or edge types). Users may apply multiple edge filters simultaneously (up to five in a single-node search) to narrow down the subnetwork. The selected filters appear as individual cards within the search field. Users can deselect a specific filter by clicking the cross on its card, or remove all active filters at once by clicking the main cross in the search box. The following are the 31 types of edges in the knowledge graph, shown with the node types they connect. Some edge types, such as related_to and part_of, connect more than one pair of node types and are therefore listed more than once.
  • Chemical—affects_activity_of—Gene
  • Chemical—affects_expression_of—Gene
  • Chemical—affects_phenotype—Phenotype
  • Chemical—causes_or_correlated_with—Disease
  • Chemical—decreases_activity_of—Gene
  • Chemical—decreases_expression_of—Gene
  • Chemical—decreases_phenotype—Phenotype
  • Chemical—increases_activity_of—Gene
  • Chemical—increases_expression_of—Gene
  • Chemical—increases_phenotype—Phenotype
  • Chemical—prototypical_stressor_of—AOP
  • Chemical—reported_to_activate—Gene
  • Chemical—reported_to_cause—DEDuCT Endpoint
  • Chemical—reported_to_inhibit—Gene
  • Chemical—structurally_similar_to—Chemical
  • DEDuCT Endpoint—related_to—Disease
  • DEDuCT Endpoint—related_to—Key Event
  • DEDuCT Endpoint—related_to—Phenotype
  • Disease—related_to—Key Event
  • Gene—acts_upstream_of—Phenotype
  • Gene—acts_upstream_of_negative_effect—Phenotype
  • Gene—acts_upstream_of_or_within—Phenotype
  • Gene—acts_upstream_of_or_within_negative_effect—Phenotype
  • Gene—acts_upstream_of_or_within_positive_effect—Phenotype
  • Gene—acts_upstream_of_positive_effect—Phenotype
  • Gene—associated_with—Disease
  • Gene—colocalizes_with—Phenotype
  • Gene—contributes_to—Phenotype
  • Gene—enables—Phenotype
  • Gene—involved_in—Phenotype
  • Gene—is_active_in—Phenotype
  • Gene—located_in—Phenotype
  • Gene—mapped_to—Key Event
  • Gene—part_of—Phenotype
  • Key Event—part_of—AOP
  • Phenotype—related_to—Key Event

2. Toolbar Overview

The toolbar located above the network view provides several utility functions:

• Neighbor Slider: Using the slider or the +/- buttons for precision, users can restrict the number of neighbors shown for a selected node. In a shortest path search, the slider instead restricts the number of paths shown. To keep the browser responsive, the maximum value of the slider is capped (2000 neighbors for a single node and 500 paths for a shortest path search). The complete data can be obtained using the Download All tool described below.
• Network/Table View: Users can switch between the default Network View and the Table View. In Table View, nodes and edges are grouped by type, paginated (10 rows per page), and searchable. Users can also export the current table as a CSV file. It is important to note that only information about the nodes visible in the network will be available in the table view.
• Export Options:
  • Save Subnetwork: Users can download the network in the present view as a ZIP folder containing node and edge tables in TSV format, or JSON format, or Cytoscape-compatible JSON format.
  • Download All (Download Icon): Users can download the complete data for the selected node(s), independent of the current view (the slider, edge and node filters, and deleted nodes are ignored). For a single node, this includes all of its neighbors; when two nodes are selected, it includes all the shortest paths between them. The data can be downloaded as a table (TSV) or as a Cytoscape-compatible JSON network.
  • Camera Tool: Users can save the current view as a PNG, JPG, or SVG. Note: Please cite DEDuCTv3.0 as specified on the Home page if using these images in publications.
• Layout & Visibility:
  • Layout Toggle: Users can switch between Force-directed, Hierarchical, or Geometric layouts.
  • Eye Tool: Users can toggle edge labels (hidden by default).
  • Legend Tool: Users can display the legend for node types and adjust its zoom level for better clarity.
  • Recenter Tool: Users can recenter the network and fit it to the available viewing area.

3. Network Interaction

• Neighbor Count Summary: Displayed just below the toolbar and above the network view, this summary shows the number of neighbors for each connected node type as x of y, where x is the number of neighbors of that type currently visible in the network view and y is the total number of neighbors of that type available for the selected node. Users can therefore quickly assess how much of the subnetwork is currently displayed versus hidden. When two nodes are selected, this summary instead displays the number of common nodes grouped by node type, along with the number of shortest paths shown out of the total number of shortest paths between the two nodes.
• Navigation: Users can click and drag nodes to reposition them in the network. Users can also click and drag on the empty space surrounding the network to move the network. Additionally, users can scroll in empty space to zoom.
• Node Information Box: Users can hover over a node to highlight its direct neighbors and reveal an Information Box split into three sections. The header displays the node type and the label of the node. The central section contains the metadata associated with the node. The fields with multiple values are shown as bullet lists. The footer contains the following three options:
  • Delete (Trash Icon): Removes the specific node from the current view.
  • Expand (Star Icon): Fetches the top 50 nearest neighbors for that node.
  • Focus (Arrow Icon): Opens the node in a new page, with a fresh network view centered on its neighborhood.
• Edge Information Box: Users can click on an edge to reveal an Information Box split into two sections. The header displays the source node type, the edge type, and the target node type. The content section shows the metadata associated with the edge. When two nodes are connected by more than one relationship, each relationship is drawn as a separate edge with its own Information Box.