Rishab Jain, then a high schooler from Portland, Oregon, won first place in Biomedical Engineering at Regeneron ISEF 2022 with a project called ICOR. Here’s the project, how it worked, and what it took to get there.
Rishab Jain, a Harvard student and ISEF winner, has a YouTube channel about science research and winning ISEF:
- How to win 1st place at ISEF: https://youtu.be/sqfzvvn2GY0
- How to design a winning poster: https://youtu.be/qWERHVs14uE
- How to come up with a research topic: https://youtu.be/AEgL_sjoE4o
- How to find a research mentor and get into a lab: https://youtu.be/tKmS4UZbfII
- Paid coaching for winning science fair: https://rishabacademy.com/stem
The Regeneron ISEF
The Regeneron ISEF brings together high school students from around the world to present their research. Projects compete across categories, and the fair awards millions of dollars in prizes and scholarships.
Rishab Jain’s Journey at Regeneron ISEF
Project Overview
In his award-winning project, Rishab applied synthetic DNA engineering with ICOR (Improving Codon Optimization with Recurrent Neural Networks). It focuses on improving codon optimization for efficient, low-cost, high-efficacy recombinant vaccine and pharmaceutical manufacturing, a real bottleneck in the industry.
The AI-Based Model
Rishab’s project is built around an AI-based model for faster, cheaper drug production using synthetic DNA engineering. The goal is to manufacture life-saving medications more efficiently and get them to patients sooner.
The Problem
Traditional codon optimization techniques, while essential for improving heterologous expression in synthetic DNA sequences, often lead to imbalanced tRNA pools and metabolic stress. This can result in cell toxicity and reduced expression, hampering the efficacy of recombinant vaccines.
The Solution
Rishab created a new codon optimization tool called ICOR, based on a recurrent neural network (RNN). ICOR was developed using a genomic dataset of Escherichia coli, a commonly used cell factory. It is designed to learn the sequential context of E. coli codon usage. The architecture, based on Rishab’s custom bidirectional long short-term memory, uses over 7,000 non-redundant, high-expression, robust E. coli genes. This allows for a better understanding of complex codon interactions.
Optimizing Genetic Codes
At the heart of the work is the optimization of genetic codes. The AI model is trained to analyze genetic sequences and select the most effective ones for drug production, streamlining the process, reducing costs, and speeding up drug development.
The Potential Impact
Rishab’s project has far-reaching implications:
- Recombinant COVID-19 Vaccines: The pandemic showed how much rapid vaccine development matters. Rishab’s model could speed up vaccine production.
- Customized Medications: Tailoring medications to a patient’s genetic makeup is the direction medicine is heading, and Rishab’s work feeds into that.
- Cost Efficiency: Traditional drug production is expensive and slow. The model could make medications more affordable and accessible.
- Performance: ICOR’s performance evaluation involves 1,481 E. coli genes and a benchmark set of 40 DNA sequences with known heterologous expression. When compared to five industry techniques, ICOR demonstrates statistically significant improvements. Notably, it achieves a 236% enhancement in real-world expression, a substantial leap over current recombinant vaccine production methods.
“I’ve been actually doing research in biomedical technology and biomedicine in general for the last four or five years now. Maybe when I was four or five, a lot of my fondest memories come from playing around on the computer.”
ISEF and Science Fair Community: Rishab Jain Academy
Rishab’s Experience
In a series of videos, Rishab walks through the project, from the role of synonymous codons to the design of ICOR. The videos give a close look at how he approached the work.
Rishab presented his research project on a deep learning-based tool to improve vaccine manufacturing. He won first place in the biomedical engineering category. His tool uses the fact that there are multiple codons for the same amino acid. By selecting specific codons, the tool can enhance protein expression, saving time and money in pharmaceutical and vaccine manufacturing.
ISEF offers a lot to participants. It’s a chance to deepen your knowledge and sharpen your research skills, meet peers from different backgrounds, and see a wide range of projects. That environment helps students broaden their perspective and improve how they solve problems.
Rishab encourages students interested in ISEF and research to take the opportunity, given how much it can shape their academic and personal growth.
“When technologies like Google Home, Amazon Alexa started coming out when I was in like sixth, seventh grade,” Jain said, “that’s when the idea of AI started to fascinate me and I got interested in AI programming.”
Recognition and Awards at Regeneron ISEF
He’s being nationally recognized for his work using AI to substantially increase the number of synthetic genes that can be produced in E. coli bacteria to create a variety of drugs. He wants to apply his algorithm to other types of cells beyond E. coli, with the ultimate goal of improving lives through his research. Jain is currently studying at Harvard, with the goal of continuing his research and earning both an MD and a PhD.
At the Regeneron ISEF 2022, he received his second Regeneron Young Scientist Award, a recognition for his work. The award came with a prize of $50,000.
”I’m interested in not only biology but medicine,” Jain said. “And that’s why I’ve been in this field of biomedical engineering and applying my skills from engineering to make an impact on a patient’s life.”
Resources to start working towards Regeneron ISEF
If Rishab’s journey is something you want to follow, he offers science fair and research coaching where he shares his winning strategies.
Whether you’re just starting or already deep into student research, keep at it. Good luck on your science journey.
Leave a Reply