Antibody Structure Prediction and Visualization with BioLuminate

Tutorial Created with Software Release: 2026-1
Topics: Antibody Design, Biologics Drug Discovery
Products Used: BioLuminate, Prime

Tutorial files

9 MB
This tutorial is written for use with a 3-button mouse with a scroll wheel.
Words found in the Glossary of Terms are shown like this: Workspacethe 3D display area in the center of the main window, where molecular structures are displayed

 

Tip: You can hover over a glossary term to display its definition. You can click on an image to expand it in the page.
Abstract:

 

This tutorial will guide you through the essential steps to predict an antibody structure, assess the structure quality of the predicted structure, and perform necessary refinements. It also covers steps to visualize and analyze interactions in antibody – antigen complexes.

 

Tutorial Content

  1. Introduction

  1. Creating Projects and Importing Structures

  1. Modeling an Antibody Structure

  1. Assessing the Structure Quality of the Predicted Structure

  1. Visualizing an Antibody – Antigen Complex

  1. Performing Interaction Analysis

  1. Conclusion and References

  1. Glossary of Terms

1. Introduction

Antibodies, also known as immunoglobulins, play a crucial role in immune response by recognizing specific antigens in our body. Each antibody is composed of four polypeptides – two heavy chains and two light chains. Both heavy and light chains consist of constant and variable domains. The constant domains (CH and CL) do not bind to antigen and are the same in all antibodies. The variable domains from both heavy and light chains (VH and VL) interact with each other to form unique antigen binding sites capable of binding to a specific antigen. The variable domains (Fvs) are further divided into framework regions (FRs) and hypervariable regions. The FRs of the VH and VL domains are highly conserved. The hypervariable regions – also known as complementarity determining regions (CDRs) – are composed of six hypervariable loops on the antibody surface (denoted as H1, H2, and H3 of the heavy variable domain; and L1, L2, and L3 of the light variable domain). Given their antigen specificity, antibodies have significant therapeutic potential.

Structure prediction of antibodies is a crucial step in the modeling, engineering, and designing novel antibodies for specific antigen targets. Modeling antibody structures typically involves combining sequence information with computational tools that predict how the CDR loops fold. While it is relatively easy to predict other CDR loops, prediction of the H3 loop is challenging as H3 loops are highly diverse in length and conformation.

Antibody structure prediction can be performed using different approaches, notably template-based homology modeling and machine learning (ML) structure prediction. The table below summarizes the respective domains of applicability for these two methods:-

Homology Modeling

ML Structure Prediction

  • This method is useful when high-homology loop templates are available.
  • This method is useful when only low-homology templates are available.
  • It is also the best way to use proprietary in-house models as templates for predicting homologous structures.
  • This approach offers a valuable alternative for predicting structures of a larger library of sequences, and can be run at a much higher throughput than the template-based method.

For more details on antibody structure prediction, please watch this video.

In this tutorial, you will learn how to:

  • model antibody structures using the template-based homology modeling approach
  • evaluate the quality of predicted structures and perform necessary refinements
  • visualize antibody – antigen complexes using presets and other options
  • perform interaction analysis to identify key residues involved in antigen binding

2. Creating Project and Importing Structures

  1. Open BioLuminate and create a new project named antibody_modeling.prj for this tutorial.
  2. Download the tutorial zip file including input files and reference outputs here: https://www.schrodinger.com/sites/default/files/s3/release/current/Tutorials/zip/antibody_prediction_visualization.zip
  3. After downloading the zip file, unzip the contents in your Working Directorythe location where files are saved for ease of access throughout the tutorial.

3. Modeling an Antibody Structure

In this section, you will predict the structure of the variable region of 1E6J structure – a murine (mouse-derived) antibody – using the Antibody Structure Prediction panel. This panel organizes structure prediction into three sequential steps – Format, Framework, and Loops. These tabs guide you to define the antibody construct, import sequences and select appropriate structural framework, and finally generate variable loop regions to build a 3D antibody model.

You will use a homology-based approach which relies on the principle that antibodies with similar sequences adopt similar three-dimensional folds. As crystal structures frequently do not exist for murine antibodies, structure prediction is a common first step in the antibody modeling process.

 

Figure 3-1. Specifying the antibody format in the Antibody Structure Prediction panel.

  1. Find and open Antibody Structure Prediction panel from Tasks.
  2. In the panel, for Build Model, choose Single.

 

First, you need to specify the desired antibody format.

 

  1. For Format, choose Fv or scFv.
    • The antibody diagram updates according to the chosen format.

 

Note: You can hover over to identify the outlined antibody regions in the diagram.

BioLuminate offers several antibody formats to match different experimental constructs. To know more, please refer to this documentation page.

 

In this tutorial, we will work with the Fv or scFv format, as our goal is to model only the variable regions of a murine antibody. Selecting this option updates the panel so that you can directly input the heavy and light variable sequences needed for homology-based structure prediction.

Figure 3-2. Importing the sequences for the heavy and light chains.

Next, you will import the sequences for the antibody regions. For this, we have provided a FASTA file with heavy and light chain sequences of 1E6J structure.

 

  1. Click on the VH (Heavy Variable) region and choose File.
  2. Navigate to FAB.fasta in the tutorial files and click Open.
    • The VH region is colored dark blue.
  3. Next, click on the VL (Light Variable) region and choose File.
  4. Navigate to FAB.fasta in the tutorial files and click Open.
    • The VL region is colored light blue.
  5. Click Framework Search at the bottom of the panel.
    • The Framework tab opens.

Figure 3-3. Running homology search.

 

Now you will specify the numbering scheme and then choose the Framework source.

 

  1. For Numbering Scheme, choose Chothia.

Note: The definition of CDR loops, ranking of frameworks, CDR loop clustering, and choice of loop templates are affected by the numbering scheme.

  1. For Framework from, choose Homology Search.
  2. Click Run.
    • This takes a few seconds.
    • The table populates with the homologs.
    • The homologs are sorted by the Composite Score, with the highest-scoring homolog selected by default.

 

Note: If two homologs have the same Composite Score, then we recommend choosing the higher-quality homolog (one with lower numerical value of PDB resolution).

Figure 3-4. Choosing the template framework.

 

For pedagogical reasons, we will use the fourth best option in this tutorial, as the generated steric clashes will help better demonstrate a potential workflow.

 

  1. Choose 2FAT row as the template for the framework from the table.
  2. Click Loop Modeling at the bottom of the panel.
    • The Loops tab opens.
    • It takes a few seconds for loop clustering to complete.
    • In contrast to other loops, which have greater than 80% similarity to loop templates, the CDR-H3 loop shows a significantly lower similarity of only 45%.

 

Note: Templates for the loops are automatically determined based on the sequence similarity of loop clusters. Additionally, you can build multiple models at once through Models to build option.

Figure 3-5. Running the structure prediction job.

  1. For Job name, type 1E6J_structure_prediction.
  2. Click Generate Loop Models.
    • This job takes ~ 5 minutes to finish.
    • Once the job is finished, banners appear and structures are added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.

 

Note: If you are not running the job yourself, the pre-generated results are included in the tutorial files. You can look at the pre-generated results by importing 1E6J_structure_prediction-out.maegz.

Figure 3-6. Visualizing the predicted antibody structure.

 

  1. Click View in Project in the banner.
    • The predicted antibody structure with a color key is shown in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
    • Blue indicates conserved residues, Cyan indicates side chain mutated residues, Red indicates predicted residues, and Maroon indicates the antibody CDR loops.
  2. Close the Antibody Structure Prediction panel.

Figure 3-7. Aligning the predicted antibody structure with the crystal structure.

 

Now you will compare the predicted structure with the crystal structure.

 

  1. Go to Files > Import Structures.
  2. Navigate to and import 1E6J_crystal_structure.maegz from the tutorial files in your Working Directorythe location where files are saved.
  3. Confirm both the predicted 1E6J antibody structure and the prepared crystal structure are included in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  4. Find and open Protein Structure Alignment panel from Tasks.
  5. For Use proteins from, choose Workspace (included entries).
  6. Click Align.
    • This takes a few seconds.
    • Once the structures are aligned, the Protein Structure Alignment Results are shown.

Figure 3-8. Viewing the Protein Structure Alignment results.

  1. Scroll to the bottom to see the Alignment Score and the RMSD values.

 

Note: The lower the Alignment Score and the RMSD, the better.

 

  1. Close the Protein Structure Alignment Results and the Protein Structure Alignment panel.

The structural alignment between the predicted antibody model and the prepared crystal structure indicates a high level of agreement. The low alignment score of 0.050 reflects a strong global structural similarity, while the RMSD of 1.113 Å suggests only minor deviations between the two structures. Such a low RMSD is consistent with an accurate backbone placement and well-preserved overall fold, with any differences likely arising from flexible regions such as loop conformations.

Now that you have predicted the structure of the 1E6J murine antibody via homology modeling, you can assess its quality compared to other structures in the Protein Data Bank.

4. Assessing the Structure Quality of the Predicted Structure

Assessing the quality and reliability of the predicted antibody model is an important prerequisite before proceeding to subsequent computational analyses, such as antibody-antigen docking. In this section, you will use the Structure Reliability Report to examine potential structural issues. Reviewing the structure reliability report helps identify the parts of the model that may require further refinement. To know in detail about the various metrics used to assess the structure quality, please refer to the Protein Reliability Report documentation.

Figure 4-1. Assessing the structure reliability of the predicted antibody structure.

  1. Includethe entry is represented in the Workspace, the circle in the In column is blue only the 1E6J predicted antibody structure in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  2. Click the Workflow icon next to 1E6J_structure_prediction-out header and choose Assess Structure Reliability.
    • The Protein Reliability Report panel opens.

Figure 4-2. Running the job to generate the structure reliability report.

You will analyze the entire structure to assess its quality and reliability.

  1. Change the Job name to prot_rel_1E6J_predicted.
  2. Click Run.
    • This job takes a few seconds.
    • A banner appears and a new entry is added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion table.

Figure 4-3. The Protein Reliability Report for the predicted antibody structure.

  1. Click Protein Reliability Report in the banner.
    • The Protein Reliability Report Panel updates to show the report.

Note: The Protein Reliability Report can also be viewed by including the entry under prot_rel_1E6J-reliability-out group in the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion and reopening the Protein Reliability Report Panel from Tasks.

  1. Optional: Click on a metric title to see the values that lie outside of the acceptable range in the adjacent table on the right.

In the protein reliability report, small green circles indicate properties that fall within a reasonable range when compared to high-quality structures in the Protein Data Bank, suggesting no potential concerns. In contrast, large red circles highlight parameters that deviate significantly from expected values and may require closer inspection.

 

When evaluating a structure, attention is given to three key outputs:-

  • Steric Clashes, which point to unfavorable atomic overlaps
  • Missing Atoms, which indicate missing side chains or residues
  • Missing Loops, which identify regions of protein that were not modeled

 

In the Protein Reliability Report for the predicted 1E6J antibody structure, the steric clashes are the potential matter of concern. Note that there is no single recommended workflow for addressing issues associated with a predicted antibody structure. While minimization through the Protein Preparation Workflow can frequently be beneficial, it is recommended to re-run the Protein Reliability Report to see if the minimization properly addressed the observed issues, or perhaps introduced new ones.

Figure 4-4. Running restrained minimization for the predicted antibody structure.

  1. Click Protein Preparation in the Favorites toolbar.
    • The Protein preparation Workflow panel opens.
  2. Click INTERACTIVE to switch to Interactive mode.
  3. Under Minimize and Delete Waters, click Settings.
  4. Under Delete Waters, make sure Distant from ligands (hets) is unchecked.
  5. Click Clean Up.
    • This job takes ~ 1 minute.
    • A restrained minimization of hydrogens and heavy atoms is performed.
    • A new Entry group is added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion with the minimized structure includedthe entry is represented in the Workspace, the circle in the In column is blue in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  6. Close the Protein Preparation Workflow panel.

Figure 4-5. The Protein Reliability Report the minimized predicted structure.

Now you will recalculate the Structure Reliability Report.

 

  1. In the Protein Reliability Report panel, change the Job name to prot_rel_1E6J_minimized.
  2. Click Run.
    • This job takes a few seconds.
    • A banner appears and a new entry is added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.
  3. Click Protein Reliability Report in the banner to see the report for the minimized structure.
    • Minimizing the antibody structure using the Protein Preparation Workflow resulted in an improvement in protein quality.
    • The steric clashes, the major cause of concern in the initial predicted structure, are now within a reasonable distance of the average PDB structure.
  4. Close the Protein Reliability Report.

Recall that the sequence similarity for CDR-H3 loop was 45%. The following steps demonstrate advanced Prime loop refinement using the Antibody Loop Modeling panel. Please note that we recommend refining the CDR loops if they are less than 40% similar to the loop template.

 

Optional: To refine the CDR loops in, you may follow these steps:-

  1. Select the 1E6J_structure_prediction-out entry group and click the Workflow icon next to it.
  2. Choose Refine Antibody Loops.
  3. Load the selected structure in the Antibody Loop Modeling panel.
  4. Choose the Numbering Scheme (in this case, Chothia).
  5. Choose the CDR loop(s) to be refined (in this case, CDR-H3).

Note: We strongly advise against running this job as it is extremely time-consuming. You can find  the pre-generated results in the tutorial files.

 

To see the pre-generated results, import 1E6J_H3_adv_loop_bld-out.maegz from the tutorial files. Additionally, to compare the results, align the refined predicted antibody structure with the prepared crystal structure (1E6J_crystal_structure.maegz in the tutorial files) and see the Alignment Score and the RMSD. Please note that these metrics are expected to improve for the refined predicted antibody structure.

5. Visualizing an Antibody – Antigen Complex

Structure visualization helps understand the structural components, spatial arrangements of atoms and connectivity of chemical bonds. Maestro provides an easy to use interface for visualizing molecules. It offers a wide range of display styles to represent molecules and customize color schemes based on different properties. In this section, you will visualize the antibody-antigen complex in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed using Preset rendering options. Further, you will also visualize the antibody-antigen interactions using the Interactions Toggle in the Workspace Configuration Toolbar.

For demonstrating the workflow in this section, we will use the 4M5Z complex structure, which consists of a 5J8 human antibody bound to hemagglutinin subunit HA1 of influenza virus. To save time, the 4M5Z structure has been prepared using the Protein Preparation Workflow and is included in the tutorial files. Please refer to the Introduction to Structure Preparation and Visualization tutorial and Best Practices for Protein Preparation before starting out with your own structure.

Figure 5-1. Applying antibody Presets.

Presets are available to quickly apply common styles to biomolecules for easy visualization. You will now visualize the complex structure using Preset rendering options.

 

  1. Go to Files > Import Structures.
  2. Navigate to and choose the proteinprep_4M5Z-out.maegz from the tutorial files in your Working Directorythe location where files are saved.
  3. Click Open.
    • The structure is added to the Entriesa simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion.
  4. Includethe entry is represented in the Workspace, the circle in the In column is blue only 4M5Z – prepared entry in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
  5. Go to Presets > Antibody > Chothia.
    • The structure is rendered with ribbons and side chains are hidden.
    • The heavy chain is colored in blue, the light chain is colored in red.
    • Constant regions are dark hues and CDR regions are dark hues.
    • The antigen is colored in green.

Figure 5-2. Selecting atoms at the antibody – antigen interface using predefined atom sets.

The Quick Select in the Selection toolbar allows you to select predefined atom sets like Vernier Zone, VH-VL interface, antibody CDRs, etc. See this documentation page for more information.

 

To visualize the interactions in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed, it is essential to display the atoms involved in the interactions.

 

  1. Under Quick Select, click the Choose item button (three dots) and choose Antibody-Antigen interface.
    • The antibody-antigen interface atoms are selected in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.

Figure 5-3. Displaying the antibody – antigen interface atoms.

  1. Go to the Style toolbox and choose Display selected atoms (open-eye button).
    • The antibody-antigen interface atoms are now visible in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.

Figure 5-4. Displaying the antibody – antigen interactions in the Workspace.

  1. Right-click the Interactions toggle and confirm that Non-covalent bonds and Pi interactions are both turned on.
  2. Change the Ligand-Receptor interactions to Antibody-Antigen for both Non-covalent bonds and Pi interactions.
    • The antibody-antigen interactions are shown as colored dashed lines in the Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.

Figure 5-5. Visualizing the antibody – antigen interactions.

Take your time to identify the CDR residues interacting with the antigen residues.

6. Performing Interaction Analysis

Protein-protein interaction analysis is important in understanding how antibodies recognize and bind their targets at the molecular level. Calculating and examining the noncovalent interactions – such as hydrogen bonds, salt bridges, hydrophobic contacts, and pi-pi interactions – helps identify the residues which play a key role in stabilizing the antibody – antigen complex. This information helps explain binding specificity and affinity, highlights important contact regions within the CDR loops, and reveals interaction patterns that may not be obvious from visual inspection alone. Such insights are valuable for rational antibody optimization, guiding mutations to improve binding, stability, or specificity, and for building confidence in predicted or modeled antibody structures.

In this section, you will use the Protein Interaction Analysis panel to identify important contacts in the 4M5Z complex structure.

Figure 6-1. Performing interaction analysis.

 

  1. Find and open Protein Interaction Analysis panel from Tasks.
  2. Under Interaction sets section, for Define sets by option, choose Chain.

 

The interactions are evaluated between two user-specified atom sets representing the antibody and the antigen.

 

  1. For Set 1, click Add and select chains H and L (of antibody).
  2. For Set 2, click Add and select chain A (of antigen).

 

Note: You can customize the distance thresholds for the interactions using the “Advanced Options” depending on the context of analysis.

 

  1. Click Calculate.
    • This calculation takes a few seconds.
    • The Interactions table populates.

Figure 6-2. Sorting the interactions by distance for easier visualization.

 

  1. Click on the Distance heading to sort the table by distance.
  2. Check the Fit on Select option.
  3. Click on a table row to visualize the corresponding interactions in the 3D Workspacethe 3D display area in the center of the main window, where molecular structures are displayed.
    • Take your time to scroll and view the interactions in the Specific Interactions column. Additionally, scroll to the right or resize the window to view other column results.
  4. Optional: Compare the calculated interactions with the findings reported in this publication.

 

Note: You can export the Interactions table as a CSV file by clicking on the Export Table option.

The calculated interactions align with the findings reported in this publication. Antibody residues such as H:ASP 100B and H:ARG 97 form hydrogen bonds and salt bridges with conserved HA residues including A:THR 136, A:ALA 137, A:ASP 190, and A:GLN 226. The interface has dense network of strong electrostatic and hydrogen-bonding interactions between acidic residues on the antibody (e.g. L:ASP 51 and L:ASP 53) and positively charged LYS residues (e.g. A: LYS 145 and A:LYS 133A) on the antigen. Further, H:PRO 100A interacts with conserved antigen residues A:TRP 153 and A:LEU 194, which form a hydrophobic pocket within the receptor binding site.

Figure 6-3. The Protein Interaction Analysis diagram.

 

  1. Click Diagram.
    • The 2D Protein Interaction Analysis Diagram opens.
  2. Click the Configure View (settings) icon and check LID Legend.
    • Follow the legend to analyze the interactions.
  3. Optional: Click the Export icon and save the image of the diagram. For visual clarity, we recommend modifying the resolution while saving the image.
  4. Close the Protein Interaction Analysis Diagram and the Protein Interaction Analysis panel.

7. Conclusion and References

In this tutorial, you learned how to model antibody structures using the template-based homology modeling approach, evaluate the quality of predicted structures and perform necessary refinements. You also learned how to visualize antibody – antigen complexes and identify key residues involved in the antigen binding through protein interaction analysis.

Following the successful prediction of the antibody structure, the subsequent steps include predicting antigen structure and antibody – antigen docking to get the complex structure, which provides atomic-level insights into the binding interface. These structural insights are then leveraged for antibody engineering, a process aimed at optimizing critical therapeutic properties such as stability, affinity, and minimized immunogenicity. The final step is developability assessment to evaluate the drug’s overall suitability for pharmaceutical manufacturing and clinical progression. For more information, please consult the Antibody Modeling learning path or further learning section below for additional resources.

8. Glossary of Terms

Entries - a simplified view of the Project Table that allows you to perform basic operations such as selection and inclusion

Included - the entry is represented in the Workspace, the circle in the In column is blue

Project Table - displays the contents of a project and is also an interface for performing operations on selected entries, viewing properties, and organizing structures and data

Recent actions - This is a list of your recent actions, which you can use to reopen a panel, displayed below the Browse row. (Right-click to delete.)

Scratch Project - a temporary project in which work is not saved, closing a scratch project removes all current work and begins a new scratch project

Selected - (1) the atoms are chosen in the Workspace. These atoms are referred to as "the selection" or "the atom selection". Workspace operations are performed on the selected atoms. (2) The entry is chosen in the Entries (and Project Table) and the row for the entry is highlighted. Project operations are performed on all selected entries

Working Directory - the location where files are saved

Workspace - the 3D display area in the center of the main window, where molecular structures are displayed