About This Project
The mining of heavy metals accounts for about 10% of global greenhouse gas emissions. Moreover, exposure to elevated levels of metals like lead, mercury, and cadmium is estimated to cause over 1.3 million premature deaths annually worldwide. In this proposal, we will leverage state-of-the-art AI to generate heavy metal protein binders that will be displayed and screened on yeast cell surfaces, facilitating efficient metal clearance for downstream bioremediation of metal-polluted sites.
Ask the ScientistsJoin The Discussion
What is the context of this research?
Presently, regions characterized by intensive industrial operations, notably mining districts and densely populated urban areas, bear the brunt of heavy metal contamination, affecting nearly 40% of the Earth's landmass (doi.org/10.1007/978-3-7643-8340-4_6). Over 3.2 billion people, predominantly in economically disadvantaged communities, endure exposure to unsafe levels of heavy metals, such as lead and cadmium, leading to severe health repercussions (doi.org/10.2478/intox-2014-0009). As such, marginalized populations disproportionately bear the health burdens of metal pollution. Our design of metal-binding proteins thus addresses an ecological crisis and strives to rectify environmental inequality, offering an accessible solution to safeguard both the environment and human well-being.
What is the significance of this project?
The design of heavy metal binders represents a unique solution for metal pollution. Current practices often rely on either broad-spectrum chelators, such as glyphosate, which can be environmentally toxic (doi.org/10.1007/s11356-017-1080-1), or heavy machinery applicable only in large-scale deployments (doi.org/10.1016/j.envres.2022.113918). Protein binders offer a more targeted and cost-effective alternative, capable of selectively sequestering heavy metals from contaminated environments via either soluble expression, immobilization on arrays, or display on eco-friendly cellular hosts. This innovation will revolutionize the bioremediation landscape, providing a scalable and environmentally benign method to combat heavy metal pollution in both developed and developing regions.
What are the goals of the project?
Our objective is to generate proteins that can effectively bind to heavy metals (i.e., lead, mercury, cadmium) with high affinity and specificity, thereby facilitating their removal from contaminated environments.
To achieve this goal, we will immediately extend our current language model architectures to generate proteins conditioned on an input metal representation. We will then screen these binders on the surface of yeast with the heavy metal as an analyte, allowing us to down-select and optimize binders with the desired properties.
Our target criteria for the final generated binders to each metal include:
-High target metal binding affinity (Kd < 10-5 M)
-High binding specificity (selectivity constant > 10^3)
-Stability across temperatures (20-80°C) and pH values (5-10)
The budget items described herein will enable both the computational design and experimental screening of heavy metal-binding proteins in our BL1 and BL2 laboratory. Additionally, travel costs will enable researchers and students to visit collaborating labs to learn appropriate techniques as well as to share results at conferences and gain feedback of our approaches. Finally, the work will be performed by graduate students who are financially covered by the department, saving significant personnel costs. With the added 10% institutional overhead ($7500), the final total is: $82,500.
Month 3: Finalize training metal-protein sequence datasets and model architectures.
Month 6: Finalize generative models, and evaluate binders for lead and cadmium.
Month 9: Optimize general yeast display expression and screening protocols for model-generated binder libraries.
Month 12: Complete the evaluation of binding proteins for additional target heavy metals. Prepare a manuscript and release code to the community. Commence collaborations with partners for scaling.
Mar 31, 2024
Curate training metal-protein sequence datasets and finalize model architectures for training. Deploy models onto our lab's 8xA100 DGX GPU system.
Jun 30, 2024
Finalize generative models, and evaluate binders for lead and cadmium (our initial heavy metal target candidates) via standard binding assays.
Sep 30, 2024
Optimize general yeast display expression and screening protocols for model-generated binder libraries. Conduct screening and analysis for conference presentations.
Dec 31, 2024
Complete the evaluation of binding proteins for additional target heavy metals. Prepare a manuscript and release code to the community. Commence collaborations with partners.
Meet the Team
I am an Assistant Professor of Biomedical Engineering and Computer Science at Duke University. Research in my Programmable Biology Group (https://www.chatterjeelab.com/) exists at the interface of computational design and experimental engineering, specifically employing artificial intelligence (AI) to generate programmable proteins for applications in genome, proteome, and cell engineering. Having completed my SB, SM, and PhD from MIT, I have engineered genome editing technologies that represent some of the broadest, safest, and most effective CRISPR enzymes to date. More recently, my established expertise in deep learning-based design has been applied to develop transcription factor-based stem cell differentiation protocols for ovarian cell types, including primordial germ cells and oocytes. My lab’s research at Duke has now extended to the emergent field of “proteome” editing, where our team leverages generative language models to design potent “guides” peptides that bind and post-translationally modify pathogenic proteins, including those implicated in genetic diseases, viral diseases, and cancer. Overall, the long-term goals of my lab are to de novo design protein-based tools for therapeutic, agricultural, and environmental applications by integrating the newest advances in generative AI with robust experimental engineering platforms.
I am also very passionate about translating my academic research to the clinic as efficiently as possible. As such, I have co-founded two companies, UbiquiTx, Inc. and Gameto Inc., which are building upon my foundational research to develop novel protein-based cancer therapeutics and fertility solutions, respectively. Most importantly, I am a firm believer in the power of conducting impactful science in an inclusive and supportive environment. As such, I have been determined and active in creating academic and community environments that welcome students of diverse backgrounds to science and engineering.
- $0Total Donations
- $0Average Donation