Decoding life itself
The convergence of technology and biology could help solve some of the world’s biggest challenges, from finding cures for diseases to providing new ways to feed an ever-growing population.
Major breakthroughs are already happening, such as DeepMind’s AlphaFold, an AI program which utilises artificial intelligence to analyse and understand proteins.
The work was recently recognised with a Nobel Prize, shared with researchers behind similar advances, and has been touted as a game-changer for drug discovery.
But that is just the beginning of the alliance between technology and biology, think Glen Gowers and Oliver Vince, co-founders of Basecamp Research.
Their firm is generating a dataset of biodiversity, and they believe we are at an inflection point in biotech, one that will lead us down some interesting paths in the next decade.
Gowers talked to TFD about how he sees the future playing out.
We packed a lab into a nine-litre box and we put it in the back of a sledge and we lived up an ice cap in Iceland for a month.
When Darwin set sail on HMS Beagle back in 1831, little did he know that the observations he would make on that journey would go on to form an entirely new theory of how life on Earth had evolved.
Now, nearly 200 years later, a similar voyage is taking place. But this next phase of our understanding of biology will happen in the digital rather than the physical realm.
It will be computers, not humans, that are best placed to decode DNA and make connections between how molecules, proteins and cells interact. And the work these powerful computers and AI models do will help solve a range of problems, from developing new ways to manufacture foods, to creating more sustainable material designs, to discovering new medicines and deepening our understanding of diseases.
It could also put us firmly on the path to personalised medicine, thinks Gowers.
“A patient could walk into a hospital room, and we could instantly diagnose them, design a drug, and work out how to produce that drug. We’re very far away from that reality, but it is not inconceivable. Just as in architecture, you don’t build loads and loads of tiny physical models and then hope the whole thing doesn’t fall down when you build the real thing. We can do the design on a computer now and have very high confidence that the building will stand up. That’s where we think we can get to with biology,” he said.
Internet of Biology
For AI models to make sense of the building blocks of life, they will need access to a whole new database. Much of the publicly available Sequence Read Archive (SRA) data – around 70% – currently comes from just five species, according to Gowers.
“That’s the equivalent of training ChatGPT on five books,” he told TFD.
To increase the information available will require a physical journey much like the one made by HMS Beagle, although, unlike Darwin’s journey, it will be collecting the unseen organisms that form the basis of life in animals, trees, and plants.
So far, Basecamp Research has collected DNA samples from around the world – from forests, rubbish dumps, and rivers. It has visited some of the most inaccessible places on Earth, including Antarctica, volcanoes, and deep-sea shipwrecks. It has done this in partnership with more than 27 countries and a range of organisations, such as the Scripps Institution of Oceanography.
And to persuade communities to take part, it has developed a unique economic model that means all the collected data is traceable back to its source, and all the communities involved in the data collection receive royalties.
“Data collection has sort of ground to a halt because people aren’t seeing these economic upsides to developing this sort of data. We are going to continuously pay back based on every piece of revenue from every use of this data. We’ve already shown that we can scale our data to an order of magnitude larger than all public data sets because incentives are aligned,” explained Gowers.
“We think we’ve done about 60% of the global biomes,” he added.
Labs in Iceland
Taking the tools of the natural world and digitising them is no easy task and while it helps that DNA sequencing has dropped hugely in price, there are other changes that need to happen, thinks Gowers.
The big genome centres are currently located in cities such as London, Boston, and Beijing.
“But most biodiversity is not in London or Boston or China, it’s in environments that don’t have genome sequencing facilities,” said Gowers.
Basecamp Research has become the first company to do fully off-grid DNA sequencing.
“We packed a lab into a nine-litre box and we put it in the back of a sledge and we lived up an ice cap in Iceland for a month. We were doing all this DNA sequencing, and it was solar-powered and infrastructure-free,” said Gowers.
While Basecamp Research’s labs are still built on the same principles of low-infrastructure requirements, they have scaled significantly in throughput and the consistency of data generated. Today, their work could end up helping identify gene editing proteins for rare diseases and cancer, or proteins that will degrade plastic faster and better.
Changes in pharma?
So what will the next decade look like for the wider biotech industry?
Investment potential is huge, but currently, Gowers feels venture capital is going into ‘quick win’ firms that might not work out in the long term.
“There are a lot of companies being set up at the moment where a new AI model gets applied to a public data set and the model is one that is also used in other fields. Putting these two things together is new, but it’s not long-term defensible.
“So, what I think we’ll see in the VC world is initial excitement in these types of companies, but less and less investment over time. Instead, the companies that are really thinking about how to build on top of a novel data set will be the most successful long term.”
The next decade will see significant steps towards a new era in biology, he thinks.
“Even small changes in how accurately we can understand biology will make huge differences in how many drugs can get to clinic or how many crops can get through trials,” he said.
“As the data sets get larger and models scale and get more intelligent, they can start to predict this more and more complex behaviour.”
AI is on the cusp not only of changing the way drugs are discovered and manufactured, but perhaps who makes them, too.
“Will the same 20 pharma companies still be the same top 20 in future? Probably not unless they can make the transition really quickly. The winners are the ones that can make that shift to becoming AI-native.”
Societal rethink
Regulation of the food and drug industries may also need a shake-up.
“We are talking about producing the same types of products that are being approved today, but we are just saying we can do more and faster. So, potentially, that is a resourcing question for regulatory agencies. If a pharma company could develop 1000 drugs simultaneously, is the agency able to approve that many at the same time? Or, if you can accurately predict how a crop is going to behave, and are able to run trials that last days as opposed to years, is an agency set up to do that?”
Life science research does not happen in a bubble, and current political instability around the world is having an inevitable knock-on effect, most notably President Trump cutting funding to some of the top US universities.
Gowers remains optimistic that the research happening in places such as Harvard and MIT will survive the current political turmoil, but he warns that the West has a rival knocking at its door.
“Biotech and life sciences in China are booming, and the advances there are absolutely remarkable. The IP being generated, the speed at which they are doing it, and the amount of funding and support available from the government are also remarkable.”
If governments around the world really want to take advantage of the discoveries biotech will offer, there may need to be a radical shift in thinking, particularly when it comes to democratising healthcare. AI won’t make this change happen, he said.
“That’s a question of whether we, as a society, choose to make these drugs priced appropriately and available to everyone.”
We love technology. Find out how we can help people fall in love with yours: https://www.wearetfd.com/
