How to Use CombFold Online
Try CombFold
Commercially Available Online Web Server
CombFold: A New Algorithm for Large Protein Complex Prediction
CombFold, a combinatorial and hierarchical assembly algorithm for predicting the structures of large protein assemblies. This method addresses the challenges of predicting large protein complexes, such as their size and the complexity of interactions between multiple subunits, by leveraging pairwise interactions predicted by AlphaFold2. CombFold achieves high accuracy, predicting 72% of complexes with a TM-score above 0.7 among the top-10 predictions in a benchmark of large assemblies, and can increase the structural coverage of experimentally solved structures by 20%.
How CombFold Works
CombFold operates in three main stages to assemble protein complexes from individual subunits:
Generation of Interactions: The process begins by applying AlphaFold-Multimer (AFM) to all possible subunit pairings and some larger subsets (3–5 subunits) to predict their interactions. This stage is the most computationally intensive part of the process, but it can be parallelized.
Unified Representation: From the predicted structures, the algorithm selects a single representative structure for each subunit and computes the transformations (rotations and translations) between interacting subunit pairs. Each transformation is given a confidence score based on AFM's predicted aligned error (PAE).
Combinatorial Assembly: Using the representative structures and pairwise transformations, the algorithm hierarchically and combinatorially assembles the entire complex. This stage exhaustively enumerates possible assembly trees, filtering out structures with steric clashes or violated distance restraints from sources like crosslinking mass spectrometry.
The final model confidence is a weighted score of the transformations used, which correlates with accuracy and helps identify high-quality predictions.
What is Tamarind Bio?
Tamarind Bio is a pioneering no-code bioinformatics platform built to democratize access to powerful computational tools for life scientists and researchers. Recognizing that many cutting-edge machine learning models are often difficult to deploy and use, Tamarind provides an intuitive, web-based environment that completely abstracts away the complexities of high-performance computing, software dependencies, and command-line interfaces.
The platform is designed provide easy access to biologists, chemists, and other researchers who may not have a background in programming or cloud infrastructure but want to run experimental models with their data. Key features include a user-friendly graphical interface for setting up and launching experiments, a robust API for integration into existing research pipelines, and an automated system for managing and scaling computational resources. By handling the technical heavy lifting, Tamarind empowers researchers to concentrate on their scientific questions and accelerate the pace of discovery.
Accelerating Discovery with CombFold on Tamarind
Using CombFold on a platform like Tamarind could accelerate the study of protein assemblies and drug discovery by:
Tackling Large Complexes: CombFold can predict complexes with up to 30 chains and 18,000 amino acids, a size that is challenging for many other methods. The ability to handle such large assemblies makes it a promising tool for exploring new therapeutic targets.
Integrating Experimental Data: The framework supports the integration of distance restraints from experimental data like crosslinking mass spectrometry, which can significantly increase the success rate of a prediction.
High-Throughput and Automation: While AFM prediction is computationally demanding, the assembly stage of CombFold is fast. By running the computationally intensive AFM predictions in a scalable cloud environment, researchers could quickly generate and screen a large number of assembly possibilities without needing to manage specialized hardware.
How to Use CombFold on Tamarind
To leverage CombFold's power, a researcher could follow this streamlined workflow on Tamarind:
Access the Platform: Begin by logging in to the tamarind.bio website.
Select CombFold: From the list of available computational models, choose the CombFold tool.
Input Subunit Sequences: Provide the amino acid sequences of the subunits in the complex. Optionally, you can also include distance restraints from experimental data.
Run Pairwise Predictions: The platform would run AlphaFold-Multimer (AFM) on all possible subunit pairings to generate initial interaction models.
Assemble the Complex: The platform would then run the CombFold assembly algorithm to build the final complex from the AFM-predicted interactions. The algorithm would rank the assembled structures based on their confidence scores.
Analyze and Validate: The output would provide a set of high-quality assembled structures that can be analyzed for their global accuracy and interface quality. You can use these structures to extend the coverage of experimentally solved structures or to identify and interpret the effects of mutations on complex formation.