Human Practices
Our iGEM project, IMPROViSeD (Integrated Modelling of Protein Complexes using Distance-Restraints and Energy-Assisted Modelling), has been shaped through an iterative process of ideation, expert feedback, and practical applications. This section outlines our engagement with stakeholders, detailing how their insights shaped the direction of our project. Our work not only aligns with scientific goals but also addresses broader societal challenges, particularly in therapeutic development.
Early Ideation and Exploratory Concepts
In the initial stages, we explored multiple ideas, collaborating with experts and stakeholders from various scientific fields. These brainstorming sessions were crucial for understanding the scope and limitations of different approaches. Although we didn’t pursue these ideas for iGEM, they helped us refine our thinking and ultimately led us to the development of IMPROViSeD.
Prediction of Metal Binding Proteins
This idea stemmed from our discussions with Prof. Nagasuma Chandra from the Department of Biochemistry at IISc, who specializes in computational biology and the study of genome-wide perturbations in diseases. Our goal was to leverage computational models to predict peptide sequences that could bind metals, with potential applications in bioremediation or metal-ion detection.
Following Prof. Chandra’s suggestion, we engaged with Dr. Deepesh Nagarajan from M.S. Ramaiah University of Applied Sciences. Dr. Nagarajan encouraged us to develop short peptides that could quench specific metals of interest. We focused on understanding metal-binding regions within proteins and explored the possibility of engineering these peptides for environmental or diagnostic applications.
However, after further analysis, we realized that this approach lacked novelty. Several similar projects had been developed in the past, and the practical impact seemed limited. Additionally, while computationally interesting, this idea did not fully align with iGEM’s emphasis on innovation and societal relevance. Therefore, we decided to move forward in a more promising direction.
Genome-wide Prediction of miRNA
Another concept involved predicting microRNAs (miRNAs) at a genome-wide level, a task inspired by discussions with stakeholders such as Sakshi and Mukund from Prof. Arun Kumar’s lab and Hemant Chandru Naik from Srimonta Gayen's Lab, from Developmental Biology and Genetics, IISc, as well as input from Prof. Swetha Ramdas at Center for Brain Research, IISc. miRNAs regulate gene expression post-transcriptionally and play crucial roles in various diseases, including cancer, making them attractive candidates for therapeutic targeting.
Our idea was to use computational models to predict miRNA sequences that could potentially target disease-related genes, providing a platform for personalized medicine or biomarker discovery. This would have required extensive computational analysis of genomic data, as well as experimental validation. However, during the brainstorming phase, we encountered several challenges, including the difficulty of validating miRNA targets without significant experimental infrastructure. Furthermore, we recognized that while miRNA prediction is an exciting area of research, our approach lacked the experimental novelty and broader integration required for a successful iGEM project.
Ion Channels
In our initial brainstorming, we also explored the role of ion channels, specifically focusing on the sodium-chloride co-transporter (NCC) located in the distal convoluted tubule (DCT) of kidneys. This protein is essential for sodium ion reabsorption, and mutations in the SLC12A3 gene lead to Gitelman syndrome, which results in renal loss of sodium and potassium, causing various health issues such as hypokalaemia and metabolic alkalosis.
To address this, we aimed to design a synthetic ion channel simulation to better understand NCC’s electrostatic potential and how SLC12A3 mutations affect its function. For this project, we received significant guidance from Prof. Debnath Pal from Center for Data Sciences, IISc and his student, Dibyajyoti Maity. Prof. Debnath helped us grasp the foundational concepts and provided practical insights through discussions and visualizations using PyMol. He guided us through the calculation of electrostatic potentials during protein conformational changes, teaching us to interpret results using Delphi software. Additionally, he assisted in preparing PDB files for molecular dynamics simulations with Gromacs, which were vital for our analysis.
Dibyajyoti, who made the software MD-Davis, which is used for calculating the electrostatic potential of the protein, energy values, profiles, etc. His experience and knowledge was very helpful for us to perform the necessary calculations for the project. Despite living in the USA, offered support through Google Meet sessions, teaching us how to use Delphi and MD-Davis effectively. His guidance allowed us to streamline our workflow and analyze the electrostatic potential of the NCC protein efficiently. This collaborative effort laid the groundwork for understanding the role of ion channels in our project.
Professor Arvind Penmatsa is a professor in the Molecular Biophysics Unit (MBU) in IISc. We met Prof. Penmatsa in the initial stages of ideation. He briefed us about the general mechanisms of working of ion transporters and the energetics of the process. Our protein shows 2 conformations, one facing inwards (cytoplasm) and the other facing outwards (lumen). A conformational change happens as the ion binds to the outward facing conformation, making it inward facing and leading to the subsequent release of the ions inwards. He gave us advice on whether we should continue a simulational study on the NCC protein and its mutations or not, and asked us also to look into the functionomics of different mutations that give rise to the Gitelman syndrome so that we have a coherent idea on how to proceed.
Experimental Design and Approach
Pivot to IMPROViSeD: A Focus on Protein Complexes
With the realization that our early ideas weren’t feasible within the constraints of the iGEM competition, we shifted our focus. Guided by our PI, Prof. Debnath Pal, and our instructor, Niladri Ranjan Das, we began developing a project centered around the structural modelling of protein complexes. This pivot was heavily influenced by ongoing discussions with Dr. Shruthi Viswanath, who has extensive experience working with integrated modelling platforms.
As we moved forward, our weekly meetings began to focus on how to integrate DREAM (Distance Restraints and Energy-Assisted Modelling) into an Integrated Modelling Platform (IMP) for protein complexes. This led us to the development of IMPROViSeD, a platform designed to model protein structures using experimental distance restraints derived from techniques such as cross-linking mass spectrometry (XLMS).
Around this time, we also sought input from Dr. Manjula Das, an expert in antibody-based therapeutics. She highlighted the therapeutic potential of understanding protein-protein interactions, especially in the context of cancer metastasis. This interaction led us to focus on the LCN2-MMP9 complex, a pair of proteins known to play a significant role in the spread of cancer. With Dr. Das’ input, we decided to demonstrate the power of our platform by applying it to the structural modelling of this complex. This decision provided a clear use case that aligned with both societal needs and the goals of iGEM.
Identification of Stake Holders
Identifying and engaging with key stakeholders is essential for ensuring the project’s relevance and broader impact.
AIIM-Interaction with iGEM ambassadors, iGEM judges and other iGEMers
Judges After the Mock Presentation
The feedback we received for our iGEM project was both constructive and encouraging. Judges commended the clarity of our presentation and saw significant potential in our platform for the structural computational biology field. They recommended expanding our outreach efforts, and example use cases to our wiki. There were calls to present more results to demonstrate project progress, consider the microenvironment's effects on distance predictions, and include other data types beyond distances. Additionally, the importance of developing a user-friendly interface for non-computational biologists was highlighted, as well as providing a demo video for easier understanding by a broader audience. Expanding industrial human practices was also encouraged to ensure the platform’s applicability to real-world use.
We took the feedback seriously and expanded our outreach efforts, engaging with additional industrial experts as key stakeholders. We added example use cases to our Proof of Concept page to illustrate the platform's capabilities. Our platform can now model using limited crosslinks, incorporating both experimental and synthetic results, and we've accounted for various molecular orientations. Additionally, we've ensured thorough documentation to make the platform user-friendly and accessible to a broader audience.
Ambassadors and other iGEMers During the Poster Presentation
This year’s iGEM Ambassadors, Ms. Srimathi Lakshminarayan, Ms. Hana Lukman, and Mr. Sara Thomas, encouraged us to broaden our stakeholder engagement and strengthen our outreach efforts, particularly for the Education special prize. They also recommended, we specify the cyber safety aspects to consider while making our project. Following their advice, we expanded our initiatives, reaching students and communities across the country to raise awareness about synthetic biology. We made a concerted effort to consult with all relevant stakeholders and incorporate their feedback into our project. We also incorporated a cyber safety section in our safety page to ensure that our platform is secure.
During our poster presentation, we showcased our work to other iGEM teams and ambassadors. We received valuable constructive feedback, among which one of the suggestions was to create a narrative with engaging characters to make our project more accessible to a wider audience. We took this advice seriously, and our Home page now reflects this approach.
Interactions with PhD students
As part of our Human Practices outreach, we engaged with PhD students specializing in structural biology to gain insights that were pivotal to our project. Roohani Basavaraj from Prof. Utpal Tatu’s Lab in the Biochemistry Department provided crucial guidance on the in-gel digestion protocol and advised us on proper protein handling during our XLMS (Cross-Linking Mass Spectrometry) workflow. We incorporated all of her advice into our wet lab protocol. Additionally, we consulted Muskaan from Dr. Shruthi Viswanath’s Lab at the National Centre for Biological Sciences, who offered valuable assistance in navigating cross-link databases, helping us refine our approach to data analysis. These interactions significantly enriched our understanding of key experimental processes.
Refining Our Approach: Meetings with Experts
As we refined our project, we engaged with various experts who helped us address technical challenges and optimize our methodology.
Dr. Alexandre Bonvin
A key figure in the development of the HADDOCK platform for modelling biomolecular complexes, Dr. Bonvin provided invaluable feedback on improving the accuracy of our structural models. His recommendations included:
-
Inclusion of Surface Atoms:
Dr. Bonvin highlighted the potential for clashes when only using C-alpha atoms for localization and suggested including surface atoms with lower bounds between them. While we initially had concerns about the large number of lower bounds this might introduce, we adapted his advice by first localizing C-alpha atoms and then refining the model by focusing on surface atoms from the protein interface.
-
Handling Protein Flexibility:
He also recommended breaking down flexible regions of the protein and performing multi-registration, a suggestion that we incorporated into our approach.
-
Crosslink Ambiguities:
Assigning experimental crosslink data to specific residue pairs can be challenging due to ambiguities. Dr. Bonvin reassured us that as long as the C-alpha distances were comparable, the proposed method would be tolerant to these ambiguities.
-
Iterative Pruning:
Finally, he advised iterating multiple times to generate different localization solutions and then pruning based on crosslink violations. This iterative process became a core aspect of our modelling workflow.
Dr. Pedro Beltrao
The professor provided critical feedback on improving our integrative protein modelling platform. Their recommendations were:
-
Time Efficiency of Guided Docking:
The professor questioned whether our platform could outperform others in terms of speed. We explained that our approach uses automated guided docking, where localization takes less than a second, and registration completes in 3-4 seconds. This allows each iteration to finish in under 10 seconds, significantly faster than existing platforms.
-
Handling Crosslink Ambiguities:
They inquired about the accuracy of our algorithm in distinguishing between crosslinks, such as residue A to B versus A to C. We confirmed that our platform has been rigorously tested to resolve such ambiguities and has consistently delivered successful results.
-
Comparison with Other Platforms:
The professor emphasized the need to demonstrate how our platform compares with others in terms of processing time. They noted that by speeding up the initial conformation formulation using crosslink data, our platform could perform more detailed energy calculations earlier in the process.
-
Avoiding Energy Minimization Artifacts:
We discussed how our team’s computational expertise allows us to minimize the chances of introducing artifacts. By using experimental evidence as a starting point rather than random structures, we reduce the risk of inaccuracies in energy minimization.
-
Escaping Local Minimums:
The professor highlighted the importance of avoiding local minimum traps, particularly in cases involving NMR data. They agreed that our method could help resolve this issue by improving the process of escaping incorrect structural interfaces.
Dr. Mahipal Ganji
The professor provided valuable guidance in securing resources within our university for the UV crosslinking experiments we initially planned.
-
Selection of Crosslinkers:
They recommended broadening our range of chemical crosslinkers to enhance accuracy, advising us to consider different crosslinkers based on the experimental context. By targeting specific amino acid residues or chemical groups, we could significantly improve the precision of our protein models.
-
Crosslinkers Like DSG:
The professor emphasized the potential of using DSG (Disuccinimidyl Glutarate), a crosslinker known for its effectiveness in capturing interactions between protein subunits. Incorporating DSG into our protocol would not only increase the accuracy of our integrative models but also strengthen their overall robustness. They encouraged us to experiment with a variety of crosslinkers to further refine the platform, potentially expanding its capabilities in future iterations.
Industry and Medical Foundation Engagement
As part of our efforts to ensure that our project had real-world applications, we reached out to both industry professionals and medical experts for their input.
ThermoFisher Scientific
We had the opportunity to present our work to a panel of experts at ThermoFisher Scientific, a leading company in scientific instrumentation and research solutions. During this meeting, the panel provided constructive feedback on how our platform could be applied to large-scale datasets, emphasizing the potential for IMPROViSeD to be used in drug discovery and other protein interaction studies. The discussions encouraged us to think about scalability and real-world applications, further refining our approach to ensure that our project could have a tangible impact beyond the academic sphere.
Majumdar Shaw Medical Foundation
We also presented our work at the Majumdar Shaw Medical Foundation, a prominent institution in cancer research and treatment. Here, we received specific feedback on the potential applications of IMPROViSeD in modelling cancer-related protein complexes. The experts we interacted with underscored the importance of accurate structural data for the development of antibody-based therapies. By focusing on the LCN2-MMP9 complex, we positioned our project to directly contribute to cancer metastasis research, with potential long-term benefits in therapeutic development.
Conclusion: A Collaborative and Impactful Project
Our project, IMPROViSeD, has been shaped by continuous stakeholder engagement, expert feedback, and real-world considerations. Through our early explorations of ideas such as metal-binding peptides and miRNA prediction, we gained a deeper understanding of computational biology and its limitations. This process led us to focus on a project with clear societal relevance—the structural modelling of protein complexes involved in cancer metastasis.
By learning from experts such as Dr. Alexandre Bonvin and Dr. Pedro Beltrao, we have built a platform that not only addresses fundamental questions in structural biology but also has the potential to improve therapeutic development for cancer treatment.
Our interactions with industry professionals at ThermoFisher Scientific and medical experts at the Majumdar Shaw Medical Foundation have further broadened the scope of our project, ensuring that IMPROViSeD can be applied to real-world problems in drug discovery and cancer therapeutics. Through this collaborative and iterative process, we believe IMPROViSeD has the potential to make a lasting impact on the scientific and medical communities.