spacer
home > ebr > winter 2022 > the rise of machine learning and what it means for biological and chemical data
PUBLICATIONS
European Biopharmaceutical Review

The Rise of Machine Learning and What it Means For Biological and Chemical Data

Over the last decade, machine learning (ML) – and in particular its subdiscipline, deep learning (DL) – has evolved from a mostly theoretical discipline to a very practical one that has permeated almost every field. This DL revolution has been fuelled by an exponential increase in data and by a matching availability of specialised hardware that can support DL algorithms leveraging this data, both while fitting the models and when using them to make predictions. The consequences of this revolution are prevalent all around us, becoming more ubiquitous every day: facial recognition and voice assistants in the smartphones we carry in our pockets, automatic surveying on industrial facilities, assisted technologies for construction, the automation of radiological evaluations for certain diseases, or the optimisation of agricultural yields.

The advent of this revolution has made pharmaceutical research the perfect ground for change. As of today, the process of drug discovery is a long and costly endeavour, with a typical time frame of 10 to 20 years until market release, and an estimated cost of over $2 billion, with an approximate failure rate of 94% (1). Subsequently, there has long been a need for tools that allow us to identify better candidates in a faster and more efficient way.

AI in the Context of Drug Discovery

The combination of data and ML has already started to show its potential as a new paradigm for drug discovery. In 2019, Insilico Medicine reported that by employing deep reinforcement learning with their model GENTRL, they accelerated the generation of lead candidates from the pharma average of 1.8 years to just 46 days (2). AstraZeneca’s Molecular AI group has been constructing an ML suite to support their operations, using models like Reinvent, a model designed for de novo drug discovery, in combination with AzDock, a tool for docking simulations, and AiZynthFinder, a model for retrosynthesis planning (3, 4).

Read full article from PDF >>

Rate this article You must be a member of the site to make a vote.  
Average rating:
0
     

There are no comments in regards to this article.

spacer
Nil Adell-Mill is a Machine Learning Engineer at Arctoris. Previously, he was part of projects on DL for formulation and immunotherapy in Novartis (Basel, Switzerland) and MyNeo (Gent, Belgium) respectively. Nil holds a Master’s degree from ETH Zurich and Zurich University, Switzerland, in Neural Systems and a Bachelor’s in biomedical engineering, with a thesis carried out at IBM Research on machine learning applied to polymer discovery.
spacer
Nil Adell-Mill
spacer
spacer
Print this page
Send to a friend
Privacy statement
News and Press Releases

OGM celebrates 60th anniversary

Leading plastic injection moulding company OGM (Owen Greenings and Mumford Ltd) is celebrating an impressive milestone this month with its 60th anniversary.
More info >>

White Papers

Pharmaceutical Companies: Outsourcing Combination Product Manufacturing

Phillips-Medisize

Combination products are therapeutics that combine two or more products (drug/device, biologics/device, biologics/drugs, or drug/device/biologics), regulated and sold as a single unit. Examples of combination products include drug-coated implantable devices, drugs packaged with delivery devices in medical kits, and drugs and devices packaged separately but intended to be used together.
More info >>

 

 

 

©2000-2011 Samedan Ltd.
Add to favourites

Print this page

Send to a friend
Privacy statement