|
Listen to this story
|
The parent company of Google, Alphabet, has announced the launch of a new company Isomorphic Labs – a commercial venture with a mission to accelerate biomedical breakthroughs and accelerate the drug discovery process. Demis Hassabis has taken over as the founder and CEO of Isomorphic Labs (and DeepMind). In his Twitter post, Demis has mentioned the launch and made it clear that the new launch will not impact DeepMind’s ongoing research works.
Biology – the branch that deals with living organisms and their vital processes, can be looked upon as an information processing system. This viewpoint argues that biology and information science may share a common underlying structure – an isomorphic mapping between the two. Hence the name Isomorphic Labs.
Why now
Last year, DeepMind made a breakthrough with AlphaFold v2 – an AI system that can predict the 3D structure of proteins directly from their amino acid sequence down to atomic accuracy, thereby solving the problem of protein folding. As these state-of-the-art technologies are getting powerful and sophisticated – the ground seems fertile to deploy such technologies to solve real-world problems.
To that end, UK-registered Isomorphic Labs is likely to use technologies coming out of the DeepMind lab to accelerate the drug discovery process with an AI-first approach.
As the pandemic hit, the focus shifted towards scientists and clinicians working in the labs to combat the disease and come out with effective vaccines. Hence, the greatest amount of private investment in 2020 was made in drugs, cancer, molecular, drug discovery, which was nearly 4.5 times higher than the amount invested in 2019, as per the latest report from Stanford University. Moreover, the drug discovery market is expected to reach $71 billion by 2025, thereby promising strong potential.
Image: Stanford
As years of research from DeepMind has begun to yield results, the company registered a first-ever profit of £43.8 million in 2020, compared to a loss of £649 million in 2019. So the time presents a golden opportunity for the team to encash their years of toil in research. Also, the team believes that a foundational use of cutting-edge computational and AI approaches can assist scientists in taking their work to the next level, accelerating the drug discovery process significantly.
The idea behind starting off this commercial venture is to make effective collaboration, wherever necessary, between the two companies. Demis also seems hopeful of partnering with pharmaceutical and biomedical companies. The collaboration will be a win-win situation as the company already hosts a world-class multidisciplinary team with expertise in the area of AI, medicinal chemistry, engineering, biophysics and biology, with a good track record in research. At the same time, large-scale biomedical companies can infuse the required funds to quickly translate research into industrial applications.
What all applications can be expected
Artificial intelligence is appealing for drug discovery because it uses the rapid and large number-crunching capabilities of the current computing technologies, such as machine learning, to compare and analyse data in the same manner that the human brain does, but in a fraction of the time. As a result, AI can be used effectively in different parts of drug discovery which includes:
- Drug design: It can help in predicting the 3D structure of proteins as done by AlphaFold 2.0, determine drug activity and drug-protein interactions.
- Polypharmacology: AI can help in designing the biospecific drug molecules and multitarget drug molecules.
- Drug repurposing: AI can help identify therapeutic targets and can also predict its new use cases.
- Chemical synthesis: AI helps in the prediction of reaction yield, design the synthetic route, and develop insights into reaction mechanisms.
- Drug screening: Here, AI can predict toxicity, the bioactivity of drugs and can further help identify and classify target cells.
The success of AI depends largely on the availability of a large amount of data. However, access to such data remains an uphill task as it incurs extra costs to a company, and the data should also be reliable and high quality to ensure accurate result prediction.
