Abzu's keynote:
"Explainable AI: Designing better molecules to become drugs."
Given by Lykke Pedersen, Abzu’s Chief Pharma Officer at the 19th Annual Meeting of the Oligonucleotide Therapeutics Society 2023.
On the conference:
OTS is an open, nonprofit forum to foster research and development of oligonucleotide therapeutics.
The Oligonucleotide Therapeutics Society (OTS) Annual Meeting is a forum for the realization of the Society’s mission and goals: To foster academic and industry-based research and development of oligonucleotide therapeutics.
Leaders in the field from across the globe, alongside students, will present on an array of topics, including chemistry, AI / machine learning, genome and RNA editing, and preclinical and clinical research.
Explainable AI: Designing better molecules to become drugs.
Oligonucleotide Therapeutics Society 2023 keynote:
How do you design molecules to become drugs? That’s a tricky question to answer, as there are many protracted and complicated hurdles to clear: A molecule first has to pass many different assays, be safe and active in vivo, and complete a final regulatory assessment before it can become testable in humans.
Today, AI is being used across many aspects of drug discovery. What if AI could explain and provide insights as to why a molecule passes each checkpoint to become a final drug?
Believing that a single AI could undertake this challenge on its own is an unrealistic utopia due to data sources (both the variety of and lack of unbiased and complete datasets). Therefore, we need to take one step at a time in order to understand the mechanisms behind certain drug properties. One very initial step is to design molecules showing activity in vitro, which can then be passed on to testing in other assays and eventually in vivo.
In this talk, we will present how AI can be used to design siRNAs and get important insights to what drives siRNA activity. Short oligonucleotides such as siRNAs can be described in many different ways, for example: Counting the occurrence and position of different nucleobases, calculating various energy parameters, and investigating positional effects from chemical modifications.
But the question remains: Which of these descriptors of siRNAs provide insights to drug properties and how are they affecting one another? Testing all potential hypotheses to determine the relationships between certain drug descriptors and their impact on a certain drug property is a cumbersome task. But such a task can be expedited by AI through in silico assays, and more importantly, the right kind of AI can perform these assays quickly while at the same time reveal the driving descriptors and their internal relationships with a mathematical model.
We used our proprietary algorithm, the QLattice®, to analyze publicly available siRNA data and uncovered that certain nucleobases at specific positions had significant influence on in vitro siRNA activity in combination with an optimal binding energy in different regions.
With explainable AI in the form of symbolic regression, we can get an improved understanding of what drives drug properties in conjunction with accurate drug property predictions. These insights can be used to design active siRNAs and potentially point to differences between siRNA activity in vitro compared to in vivo.
Our poster:
Explainable AI: Designing better molecules to become drugs.
Understanding what drives drug properties — in conjunction with accurate drug activity predictions — is a clear advantage to improving drug candidate hit rates.
Download our poster from the 19th Annual Meeting of the Oligonucleotide Therapeutics Society 2023.
We combine explainable AI methodologies to support scientists to quickly build and test new hypotheses with sufficient evidence and accuracy for critical decision making.
Download our poster:
Explainable AI: Designing better molecules to become drugs.
Entering your information does not subscribe you to marketing emails from Abzu. We just worked really hard on this research, and we want to know more about you.
