Subscribe on YouTube and never miss an interview!
Dr. Radhakrishnan Mahadevan
Dr. Radhakrishnan Mahadevan
Professor - University of Toronto

AI-Assisted Genomic Modelling: Simulations in Bioengineering

Published on

Takeaways

  1. Digital innovation, such as computer representations of biological processes, gives researchers increased capabilities to produce impactful research.  
  2. Multi-disciplinary research is key to advancing bioengineering, combining biology, robotics, machine learning, and AI.
  3. Technological development has greatly increased the capabilities of bioengineering, necessitating increased investment to maintain this upward trajectory.

Action

Canada needs to capitalize on the opportunities provided by bioengineering advances in the country by solidifying a unified plan to provide infrastructure, training, and further funding for research and development. Public-private partnerships can help to push this plan forward quickly so that we do not lag behind other countries.


What is your area of research, and what are its impacts on Canada’s biomanufacturing and bioengineering ecosystem?  

I do research in the area of metabolism. Metabolism is the process by which the food we eat, glucose specifically, is burned to make energy. We build computer representations of metabolic processes or cells to design cellular processes for different applications. One of the applications we are working on in our lab is in the area of bionylon. Bionylon is created by engineering cells to make the components that are useful for making nylon. Nylon is part of the clothes we wear or it could be part of the fabric used to make chairs and more. The idea for us is to take these pathways and put them into baker’s yeast. In theory, yeast is useful not only for baking but also for making the clothes we wear. 


You put it simply, but how much more complicated is the process? 

There are a lot of processes that go behind making yeast yield nylon. First, you have to understand the pathways, and then you have to make sure that you can optimize them in some way. That optimization uses model-based methods that are basically computer representations of these processes. 

“One of the advantages of having a computer representation of biological processes is to understand them better and test multiple hypotheses.” 

Oftentimes, some parts of biology are not very clear. There are multiple mechanisms by which the processes in biology can happen. One of the advantages of having a computer representation of biological processes is to understand them better and test multiple hypotheses. If there are multiple hypotheses underlying a biological problem, we can use a model-based framework to test them out and try to figure out which hypothesis is true. That can provide more fundamental knowledge and improve our understanding of biology, which can eventually lead to improved computer representations, designs, and applications. 


Is that an area which artificial intelligence, which Canada is quite strong in, can help? 

Absolutely. The area where we expect machine learning and digital infrastructure to make a big impact is in the design, modelling, and optimization of cellular processes. People are able to design, using these artificial intelligence (AI) and machine learning tools, better strains, proteins, enzymes, and catalysts that can catalyze the processes required to make these products. In general, the combination of AI and machine learning, along with advances in the biology side, has been able to help us rapidly make new genetic designs.  


Is this similar to the speed at which the latest COVID-19 vaccine was developed? 

From the start, when the sequence was posted, within a day or two, the sequence of the vaccine design was able to be made. Then came the rest of the processes such as the regulations of going through phase 1 and phase 2 trials before moving on to clinical trials and regulatory approvals. That was where the delay was, as opposed to being on the design side. The aspects of scale-up and manufacturing are also important given that there is a shortage of some of the ingredients that go into these vaccines.  


How would you describe Canada’s current bioengineering ecosystem, our strengths, and areas we can improve?  

Canada is very well-positioned in the bioengineering ecosystem with very strong hubs throughout the country. For example, in Montreal, Toronto, and Vancouver, there is a cluster of companies and research groups that are doing work in this area.  

“Canada is very well-positioned in the bioengineering ecosystem with very strong hubs throughout the country.” 

Separately from that, Canada also has historic strengths in automation, machine learning, and robotics. That has led to high-throughput data and methods to interpret and learn from these high-throughput data sets. Canada has strong points in all of these components, in genomics, machine learning, and in high-throughput data, robotics, and omics. 

“Canada needs a coordinated and integrated approach for building infrastructure and training for bioengineering R&D.” 

In Canada, there are also some weaknesses. Canada’s weaknesses in bioengineering are primarily on the infrastructure side. There are certain groups like the National Research Council (NRC) where there are centralized places for people to undertake scaling up, manufacturing, and more, but such hubs are not more broadly spread throughout Canada. Ideally, hubs should be in different locations in the country to bring together biology, engineering, and equipment. There are fewer examples of that here, whereas in the US and other places there are a lot more investments. For example, in the US, they have recognized that this is a need and hundreds of millions of dollars are being spent in terms of setting up a national biomanufacturing ecosystem. Canada needs a coordinated and integrated approach for building infrastructure and training for bioengineering R&D. 


What are the strengths of multidisciplinary research in biology and genomics?  

That is an important topic. In the last few years, biology has become seriously information-rich, and we are now able to do analysis at the genome scale. That is at the level of genes and usually, many organisms have thousands of genes, so every gene and organism yields thousands of data points. Such data, especially if it is generated in a high-throughput manner, lends itself to machine learning and database methods. There is a combination of biology, robotics, machine learning, and AI that is required to advance bioengineering in the future. It is definitely multidisciplinary. I am a part of BioZone, which is a multidisciplinary centre where faculty members from different areas come together and tackle important problems in sustainability. 


Why is genomics important in fighting pandemics and viruses, and how does your research intersect with this area of genomics?  

Genomics is really important for designing vaccines and diagnostics to understand differences between viruses and discovering their weak links. In such cases, we need to have the genomic information so that we can understand whether a particular virus is spreading, whether a variant of the virus is spreading, if there are mutations in the virus, and if there are other ways these mutations can allow the virus to escape antibody or vaccine treatment.  


Do you agree that investing in the bioeconomy today is essential to future progress? 

Yes. This is the time to invest in biomanufacturing and bioengineering research, primarily for two reasons. We now have unprecedented access to what is going on inside cells, in terms of genomics and our ability to characterize the concentration of genes, proteins, metabolites, and more.  

“We will be able to apply bioengineering for many problems we have not even thought of yet.” 

Furthermore, advances in the area of Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-based tools have allowed us to engineer specific aspects of the genome with pinpoint accuracy. These two events create an opportunity for us to engineer microbial communities or human cells at an unprecedented level of ability and in a targeted fashion. We will be able to apply bioengineering for many problems we have not even thought of yet. Some people are talking about information storage, where instead of storing books in a library, we can now store such information in a genome and lock it away. There is a whole slew of applications that will become possible with advances in bioengineering and CRISPR-based tools. We hope that Canada is able to take advantage by investing in this particular area at this time. 


Who and what would you pitch to make Canada a leader in bioengineering?  

I would love to have 30 seconds with the Prime Minister or the Minister of Innovation, Science, and Industry. We are at a crossroads. Given the pandemic, we have a great opportunity to drive the global transition towards sustainable societies. That push is coming and Canada has the strength in genomics, AI,  and machine learning, and it is time to bring everything together through public-private partnerships to coordinate an approach to training, research, and infrastructure in bioengineering.  

Dr. Radhakrishnan Mahadevan
Dr. Radhakrishnan Mahadevan
Professor - University of Toronto

Bio: Dr. Radhakrishnan Mahadevan is Professor at the University of Toronto Department of Chemical Engineering and Applied Chemistry. He is also a member of BioZone, which is the Centre for Applied Bioscience and Bioengineering at the universityHis research interests span systems biology, synthetic biology, metabolic engineering, gut microbiome, and microbial communities. He holds a doctorate from the University of Delaware. 

 

Organization Profile: The University of Toronto is a public research university in Toronto, Ontario, Canada. It was founded in 1827, becoming the first institution of higher learning in Upper Canada. Its department of Chemical Engineering and Applied Chemistry is one of the most research-intensive chemical engineering departments in North America, receiving around 15 million dollars for research funding in 2018 alone.