Deep Learning in Drug Development

Imagine a world where thousands of molecular properties could be determined in seconds through a machine learning algorithm. The drug development process could be shortened by months or years. This concept is conceivable through deep learning—information processing that mimics the human neural network, delving layers and layers into decision making.

As described in the Pharmaceutical Research article The Next Era: Deep Learning in Pharmaceutical Research, it is time to apply deep learning to drug research datasets “such as physicochemical property prediction, formulation prediction, absorption, distribution, metabolism, excretion and toxicity (ADME/Tox), target prediction and skin permeation, etc.”¹ Scientists who can “generate predictions in silico and test them in vitro or in vivo would also be welcomed.”¹

The AAPS Journal article Deep Learning for Drug Design: an Artificial Intelligence Paradigm for Drug Discovery in the Big Data Era highlights how deep learning will be a valuable resource as small molecule drug discovery becomes more complex.² Deep learning, however, “is designed for accuracy in very large datasets and will not be as accurate or efficient using smaller datasets.”²  The authors caution against using only deep learning to determine conclusions: “we should not restrict ourselves in the traditional predictions on biological activities, ADMET properties, or pharmacokinetic simulations, but it may also be possible to integrate all the data and information systematically and achieve a new level of [artificial intelligence] in drug discovery.”²

Deep learning could also be applied to social media or Internet search engine datasets to help determine how drugs interact with the population, such as food-drug interactions, patient compliance, drug side effects,¹ or even new uses for therapies.¹

As deep learning and artificial intelligence continue to improve across multiple industries, more and more uses will become available in pharmaceutical science. Predictive modeling may be where scientists start, but the possibilities are countless.

Read more in Pharmaceutical Research and The AAPS Journal about the deep learning concept, how other artificial ntelligence is being used, applications of deep learning in drug development, and the future of deep learning.

REFERENCES

1. Ekins S. The Next Era: Deep Learning in Pharmaceutical Research. Pharm Res. 2016;33(11):2594–2603. doi:10.1007/s11095-016-2029-7
2. Jing Y, Bian Y, Hu Z, Wang L, Xie XQ. Deep Learning for Drug Design: an Artificial Intelligence Paradigm for Drug Discovery in the Big Data Era. AAPS J. 2018;20:58. doi:10.1208/s12248-018-0210-0

Articles in AAPS PharmSciTech, The AAPS Journal, and Pharmaceutical Research are available with AAPS membership. Log in to the AAPS website and click through once on any cover to gain access to all articles within that browsing session. AAPS Open articles are freely available to read anytime at .