Introduction and Overview of ClusterControl
ClusterControl is a web-based platform built to design and predict chimeric regulatory proteins by predicting functional rewiring points. Born out of a need to simplify and accelerate the regulation of biosynthetic gene clusters for our iGEM project, ClusterControl is a comprehensive tool based on evolutionary analyses.
The primary goal of ClusterControl is to minimize the trial-and-error approach typically required in gene cluster optimization by offering predictive insights. This makes it valuable not only for our project but also for researchers working in synthetic biology, metabolic engineering, and protein design, enabling them to quickly and efficiently design chimeric regulators.
ClusterControl is split into four key components:
- The database
- The rewiring webservice
- The wiki webservice
- The annotation webservice
Figure 1: 
ClusterControl Architecture Diagram
Key Features and Functionalities
Predictive Modeling
At the core of ClusterControl is its ability to predict optimal rewiring points in regulatory proteins. By analyzing evolutionary conserved regions and evolutionary coupled amino acids, the tool allows users to assess which regions of the protein can be modified while maintaining functionality. The Disrupted Direct Information Z-scored (DDIZ) metric is a key feature that quantifies the impact of rewiring points, helping to identify those that result in minimal disruption to protein function.
User-Friendly Interface
ClusterControl is designed with usability in mind. The web interface allows users to upload protein sequences, visualize predicted rewiring points, and explore detailed analysis reports, all through an intuitive dashboard. Users can quickly see which rewiring points are likely to result in successful regulatory behavior and proceed with confidence into wet lab testing.
Figure 2: 
Web interface showing the process of visualizing DDIZ scores
Modular and Scalable Design
ClusterControl is modular and can be updated easily with new datasets or methodologies. This flexibility allows researchers to use the tool not only for current projects but also as new developments in protein design and synthetic biology arise.
Job and Resource Management
ClusterControl was designed with hardware limitations and scalability in mind and therefore employs a job queuing system to ensure the underlying hardware is not overwhelmed.
Knowledge Aggregation
The integrated regulatory database and Wiki allow for community-based knowledge base building, giving researchers the ability to give back to the community by expanding the available information on regulatory proteins of the helix-turn-helix family.
Figure 3: 
Example Wiki page showing embedding map by which protein similarity is calculated.
Figure 4: 
User explanations
Figure 5: 
Upload interface
Automatic Annotation and Enrichment
The integrated annotation webservice automatically checks uploaded sequences for compatibility, and the computed ESM-based embeddings allow for suggestions of closest related, and therefore most likely to be functional, rewiring partners, along with the ability to give an overview of which functional group a regulator belongs to. In the future, we also plan to expand our annotation services by integrating further annotation and analysis tools.
Solving Practical Problems in Gene Cluster Regulation
One of the key challenges in biosynthetic gene cluster regulation is identifying functional rewiring points in regulatory proteins. These proteins often need to be modified to achieve desired regulatory outcomes, but identifying which points can be rewired without losing functionality is a labor-intensive process.
ClusterControl streamlines this process by providing predictions based on evolutionary conservation and structural stability. It dramatically reduces the time spent on trial-and-error in the lab, allowing researchers to focus on optimizing their designs with greater efficiency. By predicting outcomes computationally, researchers can make informed decisions before conducting expensive and time-consuming wet lab experiments.
Open Source and Availability
ClusterControl is available as open-source software under the MIT License, ensuring that researchers worldwide can benefit from it. Our team has hosted the source code and documentation on iGEM’s GitLab repository, making it easy for others to download, deploy, and contribute to the project. The documentation provides step-by-step instructions on how to set up and use ClusterControl, along with examples to help new users get started.
You can access the source code for ClusterControl on GitLab.
Implementation and Technical Details
ClusterControl is built using modern web technologies:
- Backend/Webapp: Flask, a Python-based microframework that powers the web interface and handles data processing.
- Database: MongoDB, used to store protein sequences, analysis results, and user submissions.
- Deployment: Docker containers are used to package and deploy ClusterControl, ensuring scalability and easy installation in any environment.
As each container might have different requirements, each webservice is deployable as its own docker container, allowing for targeted deployment. For example, the wiki webservice requires fewer resources than the GPU-heavy annotation webservice.
The tool is designed to handle complex computations efficiently. The modular nature of ClusterControl allows developers and researchers to add new features or apply it for other protein families and datasets seamlessly.
Impact and Results
ClusterControl has had a significant impact on our iGEM project by simplifying the rewiring of regulators controlling biosynthetic gene clusters. In particular, we were able to identify functional rewiring points in the NarL-YdfI chimeric regulator, and the predicted points correlated with experimentally measured regulatory activity. This demonstrated the practical application of ClusterControl and how it can directly contribute to optimizing gene clusters.
Figure 5: 
Graph showing correlation between DDIZ scores and activity
Our team has also received feedback from external testers who have found the tool intuitive and useful for their own projects, particularly in synthetic biology and metabolic engineering.
Future Directions
While ClusterControl is already a powerful tool for gene cluster regulation, we envision expanding its capabilities. Future updates will include support for additional regulatory protein families, enhanced data visualizations, and integration with other bioinformatics tools. We also plan to incorporate more advanced machine learning algorithms to improve prediction accuracy and streamline workflows for users across various fields.
Conclusion
ClusterControl represents a significant step forward in computational tools for gene cluster regulation. Its user-friendly interface and predictive modeling capabilities make it a valuable resource for researchers in synthetic biology. By providing precise predictions of rewiring points, ClusterControl helps scientists save time and effort, making it easier to optimize gene clusters for diverse applications.
We invite researchers to explore the tool, contribute to its development, and help us push the boundaries of synthetic biology. The full source code and documentation are available on our GitLab repository.