Outline
5. Compound Property Predictions
6. Submitting Proteome Screening Jobs
7. Analyzing Results from a Proteome Screen
7.1 Exploring Compound Annotations with the Analysis Dashboard
7.2 Navigating the Putative Binders Table
7.3 Exploring the Structure of the Predicted Protein-Ligand Interaction
7.4 Uncovering Information Relating to Proteins
7.5 Using Network Graphs to Explore Systems Biology
1. Logging in
To get started with Ligand Express, go to https://ligandexpress.com/ and enter your email address and password. If you have forgotten your password, click “Reset Password” and enter your email address. You will receive an email with a link to create a new password.
2. Navigating the Dashboard
After logging in, you will be directed to the Dashboard. The first time you log in you will be prompted to take a tour of Ligand Express. This tour can be restarted at any time by selecting the “Tour Dashboard” option from the dropdown under your name in the upper right corner of the screen.
The Dashboard gives you rapid access to Recent Libraries, Compounds, and Proteome Screening jobs via the orange, yellow, and green-coloured tiles respectively. If you have admin privileges, you will also be able to add new users via the cyan-coloured Add New User tile.
Your ADMET Credits (left-most icon) and Proteome Screening Credits (middle icon) are displayed in the upper right corner of the screen. Each ADMET Credit allows you to upload 1 molecule. For this molecule, you can perform unlimited POEM model predictions with both Cyclica-provided models and user-derived models from uploaded measurements data. Proteome Screening can only be performed on molecules already uploaded to Ligand Express using Proteome Screening credits, where 1 credit allows you to run 1 Proteome Screen.
Notifications regarding completion of Proteome Screen jobs can be viewed by clicking the bell icon.
3. Creating a New Library
Libraries allow you to group compounds and run computational experiments on them. To create a new library, click on the “Add New” button on the orange Dashboard tile or the "Add New" link under the LIBRARIES heading in the left-hand sidebar. Enter a name for the library and an optional (but recommended) description, then click the Submit button.
Once created, you can add new or existing compounds to the library by clicking on the respective buttons.
4. Adding New Compounds
The Compound Submission page can be accessed by:
clicking the “Add New Compounds to Library” button on a newly created library;
clicking the “Add New Compounds” button in an existing library;
or by clicking “Submit” under the COMPOUNDS heading in the left-hand sidebar.
After clicking the “Click to add a compound” box, a compound submission box will appear.
To add individual compounds, use the Smiles field and supply a unique name for the molecule.
To add individual or multiple molecules, upload a standard SDF or MOL2 file, or a comma/tab separated CSV file containing a unique name and SMILES per line.
Click the “Add” button to add the molecules to your selected library, then click “Go to Library” to navigate to the library.
5. Compound Property Predictions
In addition to organizing your compounds, Libraries can also be used to evaluate their properties. To compute properties for compounds:
- Navigate to the library where the compounds are stored.
- Click the “Heatmap Properties” button and select the properties you wish to compute from Cyclica’s in-house models or from user-submitted models.
- Check the boxes next to the models you wish to compute properties for and click the "Update" button.
You can remove already-computed properties from the library view by deselect the corresponding checkboxes in the Library Properties window.
The selected properties will be computed. Please note that there may be longer run-times when submitting more than 10 compounds at a time. Once the calculations have been completed, the properties will be displayed in the heatmap view on the library page. You can export these properties by clicking the “Download Heatmap as CSV” button in the top left of the heatmap.
Properties can be sorted or filtered by selecting the icon in the column headers. To filter, move the slider to the threshold(s) you are interested in and click “Apply Filter”. To sort molecules by the inverse of the property, select the “Invert selection” checkbox and click “Sort by Property”.
6. Submitting Proteome Screening Jobs
The Proteome Screening feature allows you to screen compounds against structurally characterised-protein pockets to predict binding interactions. To launch a Proteome Screen for a molecule or a set of molecules, navigate to the library where the molecules are stored. Click the grey Proteome Screening icon () beside the name of the molecule you want to send for screening.
Multiple compounds can be sent for Proteome Screening by clicking the checkboxes to the left of compound names and then clicking the green “Proteome Screening” button. The icons beside compound names will turn yellow while calculations are in progress and green when complete.
7. Analyzing Results from a Proteome Screen
To view the results of a Proteome Screen, navigate to the library page and click on the green Proteome Screening icon () of compounds with completed Proteome Screening jobs. Alternatively, click “Proteome Screens” under the PROTEOMICS heading in the sidebar to see a list of all Proteome Screening jobs. Click on “Launch” in the row of the compound you wish to analyze to launch Proteomics Analysis in a new tab.
The first time you launch Proteomics Analysis you will be prompted to take a tour. This tour can be restarted at any time by selecting the “Tour Analysis” option from the dropdown under your name in the upper right corner of the screen.
7.1 Exploring Compound Annotations with the Analysis Dashboard
Opening the results of a Proteome Screen will take you to the Analysis Dashboard screen in Proteomics Analysis. The Molecule Panel to the top left of the screen displays the name, SMILES string, and chemical structure of the query ligand. Click on the ‘copy’ icon after the SMILES string to copy it to your clipboard.
Below the Molecule Panel is the Toxicity Panel. This section contains our index of 63 proteins with established links to adverse side effects. The table shows the rank percentile of the protein and the similarity of the top known binder/assayed compound to the screened compound. Clicking the arrow at the start of the row will expand it to show the details of the side effect associated with that protein.
The Enrichment Panel contains a sunburst graph showing all enriched annotation terms with p < 0.05. The type of annotation can be selected from the top left dropdown menu. The color of each slice represents the -log p of the Mann-Whiteny U test statistic for the associated annotation. The size of each slice is determined by the p-value of the annotation in conjunction with those of its children when applicable. For Pathways, the graph will not display any annotations that are more than 4 levels from the original root.
Hovering your mouse over a slice will display additional details, including its name and ID. Clicking on a slice will zoom in on that slice, making it the new root. This will also select the corresponding entry in the right hand table. Click on the root to return to the previous level.
The enrichment table to the right of the graph shares a data source with the graph. In addition to the -log p shown in the graph, this table includes two additional columns: Total is the total number of screened proteins bearing the annotation, and Effect Size is the average rank percentile of those proteins. Selecting an annotation here or on the graph will populate the Proteins Involved section below, which contains a list of proteins bearing the annotation, sorted by rank percentile. Clicking on a protein name will take you to the Protein Details page for more information.
7.2 Navigating the Putative Binders Table
The Putative Protein Binder Table is a ranked list of the proteins putatively predicted to interact with your ligand that can be viewed on the left side of the Protein Viewer and Protein Details sections. The table can be expanded by clicking the Expanded Putative Protein Binder Table icon in the left-hand sidebar.
The table columns can be sorted in ascending or descending order by clicking on the column header once or twice respectively. Proteins can be favourited by clicking the star (☆) beside their Rank Percentile. To sort the table by favourited proteins, click on the star in the table header (★).
Additional columns can be introduced by clicking on the three vertical circles (⋮) in the top left corner of the table. To download the results, click the download button at the bottom of the table. A box will appear with a slider that will allow you to select the number of proteins that you wish to download.
The table can be filtered by a variety of criteria. When adding multiple filters, you can toggle between “AND” and ”OR” operators by clicking on the operator text. The “NOT” operator can be toggled by clicking the filter pill, changing the colour from green to pink. To exclude the term from the filter, click on the pill again to turn it transparent. Clicking the pill a third time will restore it to its original state.
For more complex queries, you can click and drag a filter term outside the brackets to create a second set of brackets. To remove the entire filter, click the trash icon to the left of the filters.
7.3 Exploring the Structure of the Predicted Protein-Ligand Interaction
The site of the predicted interaction can be displayed using the Protein Viewer. Protein residues in the pocket at the site of interaction are displayed in white. You can toggle between viewing the entire protein or just the pocket residues using the “Protein View” and “Pocket View” buttons respectively. The reference ligand used to make the prediction can be visualized by clicking the “Show/Hide Reference Ligand” button.
Click the "i" icon next to "Protein Viewer" for in-depth instructions on how to navigate the 3D model.
It is possible to dock your ligand into the highlighted protein pocket. Note: Docking does not influence the predicted interaction, it is merely for visualization purposes.
If you have not docked your ligand into a pocket previously, click the “Upload Ligand” button. You will be prompted to either auto-generate a structure for your ligand or to upload a custom 3D coordinate file (SDF or MOL2 format).
Once coordinates exist, the button will change to “Dock Ligand”. Click this and confirm that you wish to dock your ligand. The “Dock Ligand” button will then change to “Docking in progress”. When complete, the ligand will automatically be visible in the pocket. You can toggle it’s visibility by clicking the “Hide/Show Docked Ligand” button.
Information on Single Nucleotide Variants (SNVs) for the visualized protein can be displayed by clicking on the Variants sidebar on the right side of the page to expand it. This will reveal a table containing information regarding residue mutations, frequency of observance and predicted deleteriousness metrics. Variants are ordered by proximity to the ligand and can be highlighted on the structure by clicking on the row in the table. The position of the SNV in the PDB relative to the canonical protein sequence is displayed and highlighted below the protein in the Alignment bar. Adding columns, sorting, and filtering the table can be achieved as described for the Protein Putative Binder table (see 7.1 above).
7.4 Uncovering Information Relating to Proteins
You can view additional information relating to the selected protein by clicking the Protein Details icon in the left-hand sidebar. Uniprot descriptions and annotations are displayed in the main (top left) panel as well as in the “Extra Details” tab. Associations between the protein and diseases can be viewed in the “Diseases” tab. Articles supporting these interactions can be viewed by clicking on the number in the “Supporting Articles” column.
The "Assayed Compounds" tab displays all the compounds that have been measured against the selected protein. In this panel, you can view source database and IDs for the assayed compounds, Tanimoto similarity to the screened compound, and confidence measures. You can also view the structure of the compound to the right of the panel and supporting articles by clicking the "View supporting PubMed article(s)" button below the compound structure.
The “Abundance in Tissue” tab displays information about the abundance of the selected protein in tissues and cells. The relative abundance (in PPM) can be displayed as a chart, graph, or anatomical diagram. You can toggle between “System View” and “Graph View” using the switch in the upper left of the tab, or between adult and fetal data using the radio buttons. Hovering over an organ or cell in System View will display the name of the tissue/cell, abundance in tissue/cell, and abundance relative to other proteins in the tissue/cell (abundance percentile).
The modulatory effect of the query ligand on the selected protein can be viewed in the “Effect Prediction” tab. If no calculation has been previously performed, click the “Run Effect Prediction” button to compute the predicted effect. Predicted interaction types include "Activator", "Inhibitor" or "Unknown", and the four molecules that most resemble the query molecule are displayed for each uniref level of prediction. Uniref levels make use of compounds known to interact with the "Exact" protein, as well as with proteins having 100, 90 and 50% sequence similarities. Note that some proteins will not be able to compute an effect due to lack of data to create an accurate model.
The "Pathways" tab displays all the Reactome pathways that include the selected protein. The pathways are organized according to the pathway hierarchy to the left of the panel, and the pathway viewer is displayed to the right. The viewer is fully interactive and can be used to further explore the pathways.
7.5 Using Network Graphs to Explore Systems Biology
The results from a Proteome Screen can be displayed graphically by clicking the Network Analysis icon in the left-hand sidebar. Click on the "Apply Graph Type" dropdown in the upper left hand corner to display a visualization that annotates proteins with drug and/or cellular location information. The data can further be explored using custom filters for specific diseases, keywords, gene ontologies or locations by clicking the “ADD FILTER” text above the predefined graphs dropdown. Filters work similarly to those in the Putative Binders Table (see section 7.1 above) where a term can be excluded by clicking on the green filter pill to turn it pink.
Click on the Context Panel tab to the right of the graph to expand a dynamic panel with supplementary information. Clicking on a graph node populates the panel with additional details for the selected protein, disease, or drug, while clicking on a link displays information for both nodes and the relationship between them.
To access literature supporting interactions with proteins, hover over the link between nodes and click the document icon. Supporting literature can also be accessed in the Context Panel when a link is selected by clicking on the number next to “Pubmed Articles”. Clicking on a protein node will allow you to favourite a protein by toggling the star (☆). Favourited proteins are coloured yellow and will be updated in the Putative Binder Table. Proteins favourited in the Putative Binder table will also be favourited in the Network Graphs.
The number of proteins displayed can be changed by clicking the “SETTINGS” button and adjusting the slider. Compounds known to bind proteins can be added by checking the “Show Compounds Known to Interact with Proteins” option under the “Optional Nodes” heading. You can also toggle preferential viewing of favourited proteins by checking the “Show All Pinned Proteins”.
Additional Network Graph tabs can be created by clicking the + symbol in the upper left corner of the page, near the graph name. Graph tabs can be renamed by double clicking the header text, and the graph itself can be exported to PNG by clicking the lower right hand icon.
Further reading: