Learning Path: Antibody Modeling

Antibodies are a major class of biologics with broad therapeutic applications. Computational modeling plays an important role in different stages of antibody discovery and optimization, spanning from initial structure modeling and refinement to final developability assessment. This learning path provides an overview of Schrödinger tools and workflows for modeling antibodies, with a focus on choosing appropriate approaches based on available information and research goals.

Online courseIntroduction to computational antibody engineeringThe online course provides an overview of structure-based workflows for assessing and improving the stability, affinity, developability, and 'humanness' of antibody-therapeutics.

Sequence Analysis

The first step is gathering the available relevant sequence data. Once you have the sequence data, you can perform sequence-based analysis that helps annotate antibody sequences by identifying frameworks, CDR regions, and germline origins. It also helps to flag sequence features that may warrant closer inspection in downstream modeling.

TutorialIntroduction to Bioluminate and the Multiple Sequence Viewer(MSV) for the basics
DocumentationSequence Alignment Optionsin the MSV
DocumentationIdentify Family Features
DocumentationProtein Sequence Descriptors

Do you have a structure of the antibody-antigen complex of interest?

If you already have a structure of an antibody-antigen complex, you will need to prepare and refine the structure for downstream tasks.

For which part of the system do you want to predict a structure?

There are two approaches to predict structures – homology modeling and machine learning-based (ML) structure prediction. Homology modeling uses closely related proteins to map structure motifs to a sequence, which is only possible if structures of homologs with sufficiently high sequence identity are available. ML structure prediction uses the similarity to known structures as well, but in a much more holistic fashion. It can be useful to use both approaches and compare the results. Note that the following steps are not exclusive -- you may need to predict structures for both proteins separately, refine them, and then bring them together to obtain the complex structure.

Antibody Structure Prediction

You can predict antibody structures by classical template-based homology modeling or by machine learning approach using the Antibody Structure Prediction panel.

To learn more about the available methods, you can read more in following publications:

TutorialAntibody Structure Prediction and Visualization with BioLuminatetutorial guides you through the steps to predict antibody structures via homology modeling approach and make necessary structural refinements.
VideoAntibody Structure Predictionvideo guides you through different strategies for antibody structure prediction.
PublicationAntibody structure determination using a combination of homology modeling, energy-based refinement, and loop prediction
PublicationAb initio structure prediction of the antibody hypervariable H3 loop
PublicationImmuneBuilder: Deep-Learning models for predicting the structures of immune proteins

Antigen Structure Prediction

You can predict antigen structures by classical template-based homology modeling or by machine learning approach.

Learning PathComputational Structure PredictionA general introduction to computational structure prediction methods for protein targets.

Antibody-Antigen Complex Structure Prediction

TutorialAntibody – Antigen Docking with PIPERA step-by-step guide for antibody-antigen docking.
DocumentationProtein-Protein Docking with PIPER
DocumentationPreparation of Structures for Protein-Protein Docking
PublicationPIPER: an FFT-based protein docking program with pairwise potentialsDescribes the PIPER docking program

Structure Quality Assessment and Refinement

Assessing the structure quality of the predicted structure is an important prerequisite before proceeding to subsequent computational analysis, such as antibody-antigen docking.

While structure quality assessment follows common principles for all proteins, refinement strategies often differ for antibodies and antigens due to their distinct structural features. Antibodies have conserved frameworks with highly variable CDR loops (especially CDR-H3), which often require targeted loop refinement. Antigens can range from small, rigid domains to large, flexible or multi-domain proteins, and their refinement typically focuses on overall fold quality and local geometry.

TutorialIntroduction to Structure Preparation and VisualizationShows the workflow for a generic protein.
DocumentationProtein Reliability Reportgives a graphical representation of various metrics that indicate reliability or quality of protein structures
Quick Reference SheetInterpreting the Protein Reliability Report
DocumentationBest Practices for Protein Preparation
DocumentationAntibody Loop Modelingfor refining CDR loop positions
DocumentationRefining Protein Structuresfor refining antigen structures

Antibody Engineering

Antibody engineering covers computational and experimental strategies used to design, modify, and optimize antibodies for improved therapeutic performance. This includes improving antigen binding, ensuring structural stability, minimizing developability risks, and tailoring antibody formats and functions for development. Following is a list of some generic resources that introduce you to the protein design tools to introduce mutations and predict their impact on the affinity of the complex:

Additionally, if your antibody structure is derived from a non-human source, you can perform antibody humanization via CDR grafting or residue scanning.

Online courseIntroduction to computational antibody engineeringThe online course provides an overview of structure-based workflows for assessing and improving the stability, affinity, developability, and 'humanness' of antibody-therapeutics.

Optimizing Stability and Affinity

VideoNext-generation protein engineering with physics: Predicting affinity, stability, selectivity, and pH-dependent bindingoverview webinar

MM-GBSA

TutorialImproving Antibody Stability/Affinity Using MM-GBSA Residue ScanningMM-GBSA Residue Scanning allows you to identify mutation hotspots (via alanine scanning) and improve the stability and affinity via affinity maturation/CDR mutagenesis.
Quick Reference SheetMM-GBSA Residue Scanning

FRS

FEP+ Residue Scanning (FRS) is an efficient method to triage promising mutants from MM-GBSA residue scanning before performing full protein FEP+ calculations.

TutorialIdentifying impactful mutations using FEP+ residue scanning

Protein FEP+

TutorialImproving the Thermostability of T4 Lysozyme Using Protein FEP+ Guided Design
TutorialValidating a Protein Free Energy Perturbation Model for Thermostability Predictions for Single Point Mutations

Humanization

TutorialHumanizing Antibody Structures with BioLuminate
TutorialAntibody Humanization and Back Mutationvideo
DocumentationAntibody Humanization: CDR Grafting
DocumentationAntibody Humanization: Residue Scanning

Other

TutorialDisulfide Bond Engineering

Developability Assessment

Identifying and designing out liabilities before moving to the experimental stage can significantly reduce costs and speed up project timelines.

TutorialLiability Analysis for BiologicsThis tutorial highlights a few different tools and techniques for identifying liabilities.
DocumentationProtein Surface Analyzercan identify both aggregation or reactivity hotspots.
DocumentationReactive Protein Residues Panelscans the solvent-exposed residues for reactive sites.
TutorialProtein pKa Prediction with Constant pH Molecular DynamicsThe most accurate method to determine how your protein reacts to pH changes.
DocumentationProtein Titration Curve PanelA faster, less accurate method to predict pH-dependent protein behavior.

Conclusion and Next Steps

This concludes the overview of Schrödinger tools and workflows which have been validated for use on or with antibodies and antibody-antigen complexes. For more in-depth guidance on particular scientific workflows, consult our other learning resources.

Learning PathAll Learning Paths
TutorialAll Tutorials
Online courseSchrödinger Online Certification Courses

Workflow