Rounak Dubey, MBBS, MD
Assistant Professor, Transfusion Medicine
All India Institute of Medical Sciences, Nagpur
"The oldest and strongest emotion of mankind is fear, and the oldest and strongest kind of fear is fear of the unknown" - H.P. Lovecraft
The revolutionary wave of artificial intelligence (AI) and machine learning (ML) continues to make an indelible mark on every field, including Cell and Gene Therapy (CGT). As we stand on the threshold of a new era filled with limitless possibilities, some lingering apprehensions accompany this transformative journey. Before we discuss some recent AI-driven developments, like the introduction of TIGER (Targeted Inhibition of Gene Expression via guide RNA design), maybe we can pause and reflect on these doubts and apprehensions.
For over a century, popular perceptions have been influenced by contemporary literature and movies that depict a future where machines and AI could potentially take over the world. Recently, even some prominent figures in the field of AI have called for a temporary pause in the development of more advanced models (1). Similar skepticism once surrounded CGT during its early days, but thanks to the dedicated efforts of various regulatory and scientific bodies, significant strides have been made in streamlining its progress as we move towards the future. The joint Amicus Curiae (a legal document supplied to a court of law) by ISCT and ISCR, submitted on 02 June 2023, is a recent example of these efforts (2). Looking back through the lens of retrospection, we have come a long way from the days the word cloning captured the news headlines, mainly for the wrong reasons. The potential for CGT by integrating Artificial Intelligence is intriguing, but only time can reveal how things unfold.
AI and ML models can enhance speed and accuracy in selecting the appropriate target for therapeutic success and help with the personalized approach. One recent development, published in Nature Biotechnology, introduces TIGER (Targeted Inhibition of Gene Expression via guide RNA design), an advanced AI-backed deep learning model (3). TIGER has been described as the first tool for predicting both on and off-target activity of RNA-targeting CRISPRs. It can do so by comparing the predictions made by the deep learning model and laboratory tests in human cells (4).
A detailed report by McKinsey & Company, titled How AI can accelerate R&D for cell and gene therapies, has projected that by 2030 cell therapy is expected to be the third-largest segment across all modalities in oncology, behind antibodies and small molecules (5). AI can facilitate the development of these therapies in almost every step, including target identification, payload design optimization, translational and clinical development, and complete digitization. Naturally, there will be concerns about the impact of these AI advancements on jobs in all these areas as well.
The impact of AI and ML has extended beyond just therapy development. In a recently published paper in Frontiers in Medicine, ML algorithms to identify and suggest relevant cell and gene therapy regulations proved much more efficient, cost-effective, and accurate compared to manual searches. With continuous system learning, the model could process approximately 9,000 regulations/day while 3-4 subject matter experts used to review around 115 regulations in the same time equivalent (6). There is still a big question mark on the reliability and authenticity of the results generated by AI, which may improve as these models evolve. Most major scientific publishers have advised to use AI cautiously and stated that these technologies should only be used with human oversight and control (7).
This could be merely the tip of the iceberg, and it is reasonable to anticipate additional AI-powered CGT advancements in the near future. The potential for CGT combined with the integration of AI is genuinely fascinating. As we adapt to this technological transformation, we must prioritize upskilling and reskilling the workforce to harness the full potential of AI and ensure a sustainable future. On a lighter closing note, AI tools were of great help in editing and reframing some of the ideas in this article. I am cautiously optimistic that it won't replace contributing editors in the days to come!
References:
- https://www.weforum.org/agenda/2023/03/ai-leaders-call-for-pause-plus-other-ai-stories-to-read-this-month
- https://www.isct-unprovencellulartherapies.org
- Wessels, Hans-Hermann, et al. "Prediction of on-target and off-target activity of CRISPR–Cas13d guide RNAs using deep learning." Nature Biotechnology (2023): 1-10.
- https://neurosciencenews.com/ai-crispr-gene-therapy-23573/
- https://www.mckinsey.com/industries/life-sciences/our-insights/how-ai-can-accelerate-r-and-d-for-cell-and-gene-therapies
- Schaut, W., Shrivastav, A., Ramakrishnan, S., & Bowden, R. (2023). Search, identification, and curation of cell and gene therapy product regulations using augmented intelligent systems. Frontiers in Medicine, 10, 1072767.
- https://www.thelancet.com/pb-assets/Lancet/authors/tl-info-for-authors-1686637127383.pdf
#fromtheeditors
#feature4
#30.8