Infographic – AI for drug discovery
By Katerina Mansour - 08 November 2022
The challenges behind the development of new medicines are numerous and have historically been difficult to overcome. While the general public hasn’t always been fully aware of these challenges, the topic was thoroughly discussed during the height of the pandemic, when pharmaceutical companies worldwide took to the development of a Covid-19 vaccine. At that time we were all reminded of the significant funding and the years of research typically required to develop pharmaceutical products. Yet, at that time we were also reminded of just how problematic it is that the development of new pharmaceutical products takes such lengthy times. Indeed, in situations like pandemics or outbreaks, time is of the essence. To tackle some of the core difficulties the pharmaceutical industry faces, a growing number of startups have been developing AI-based solutions. Indeed, AI for drug discovery has been a hot topic for a few years now.
The benefits of AI for drug discovery
The research and development behind the creation of a new drug costs on average $1-3 billion and can last 11 to 18 years. Furthermore, success is far from guaranteed. Indeed, the failure rate for drugs making it to commercialisation is 90%. This failure is typically associated with a lack of effect or too many side effects when tested on humans. Research also shows that between 2006 and 2015, only 9.6% of all new compounds being investigated actually made it to market. Furthermore, so much has yet to be explored when it comes to the human body. Data has shown that 40% of the human genome has yet to be explored and that only 10% of human proteins have been identified as druggable.
As a result, AI is increasingly seen as a solution to explore the unexplored and make drug discovery both faster and more successful. So, what are the key benefits outlined for the use of AI for drug discovery?
- Tapping into unstructured data more efficiently
- Aggregating data worldwide faster for better insights
- Better predicting which drugs will fail clinical trials
- Predicting the efficacy and safety of a drug
- Improving the selection of candidates for clinical trials
- Predicting new small molecules with desirable properties
- Eliminating subjective biases
Many of these benefits arise from the fact that AI is able to look at large quantities of data and discern conclusions or patterns that humans would typically miss. Indeed, AI is able to analyse and extract meaningful information from large sums of data at a speed humans would be unable to match. This analysis can help move through the many steps that come with drug development much faster, but also with more predictive abilities. Based on its analyses, AI is able to identify probable outcomes and thus help influence key decisions early-on to reduce failure rates.
However, there are also many challenges worth noting when it comes to using AI for drug discovery. Indeed, access to quality data remains an issue, and algorithm bias is also a potential risk. Furthermore, it’s important to keep in mind that anything resulting from AI must be verified by scientists for accuracy.
So far this year, Crunchbase data shows startups in drug discovery have raised more than $1.4 billion. Overall, many startups have gained momentum in this space and signed prolific deals with large pharmaceutical companies. One example is Sanofi’s recent $1.2 billion deal with startup Atomwise, to research small molecules aimed at up to five drug targets. In fact, the number of strategic partnerships for AI-based drug discovery has grown from 10 in 2015 to 105 in 2021. Indeed, it seems clear that in the coming years, startups will continue to play a key role in helping pharmaceutical companies leverage AI for their drug development.
Learn more about AI for drug discovery in our infographic below: