Deploying AI in drug discovery

Deploying AI in drug discovery

AI healthcare upstart Aladdin continues to develop best-in-class AI technologies to advance and accelerate new breakthroughs in drug discovery.

In October 2020, UK-based healthcare technology developer Aladdin successfully released its commercial proprietary Artificial Intelligence Drug Discovery (AIDD) platform, aimed at accelerating the drug development process. The automated machine-learning platform pipeline covers all the main stages of early drug discovery:

  • Virtual screening;
  • Hit to lead, also known as lead generation;
  • Lead optimisation; and
  • The pre-clinical phase.

With its proprietary platform, Aladdin has already identified and optimised multiple small molecule drug compounds with high potential for age-related diseases, including Alzheimer’s disease, Parkinson’s disease, and SARS-CoV-2 (COVID-19). Prior to the launch of its AIDD platform, Aladdin was able to leverage its Knowledge Graph to identify that the steroid dexamethasone could be used to relieve COVID-19 symptoms in March 2020, before anyone else. Dexamethasone was then reported by the BBC as a potential drug candidate for treating COVID-19 in June. The results have been published in the leading AI conference, BIOKDD.

Screenshot of the platform at work

AI drug discovery

The discovery and development process for a new drug is extremely long, expensive and challenging; and takes an average of 10 to 15 years. The average success rate, which is only 10% for each drug discovery process, is extremely low. Due to the long development period and low success rate, the development cost for a single drug has risen significantly to $2.6bn. The commercial returns for pharmaceutical companies have dropped from 10% in 2008 to around 2% in 2018. Based on the research report from Deloitte, AI-assisted drug discovery is the most promising solution for this dilemma.

The early discovery phase of the drug discovery process represents the largest portion (33%) of its total development cost. The recent breakthroughs in AI have demonstrated its potential for the pharmaceutical industry. By embedding data into a high-dimensional space and extracting key relationships, AI provides innovative solutions for all stages of early drug discovery. The market size of AI-enabled drug discovery is projected to reach $1.43bn by 2024, with an annual increase of 40.8%.

Aladdin AI Platform

Aladdin’s AI platform, by means of its groundbreaking components such as ligand- and structure-based drug design using AI, can significantly speed up the process of drug development and optimise new drug compounds.

Through this platform, we can assist pharmaceutical companies to reduce the research and development cycle by 40% to 50%, save more than 50% on the development costs and increase the success rate by more than 20%.


The platform comprises four core functions, which cover all aspects of early stage drug discovery. These functions are:

  • AI compound design;
  • AI virtual screening;
  • AI property prediction; and
  • AI synthetic path design.

We have already seen multiple successful use cases via co-operation with big pharma and institutions. All the functions have significantly outperformed those of other commercial applicants.

Aladdin AIDD Platform

Main functions

  • AI virtual compound design;
  • AI drug synthesis prediction;
  • AI drug virtual screening; and
  • AI macromolecular docking.

Algorithm advantages

  • Algorithm validated by multiple AI competitions; and
  • Multiple successful business and academic partnerships.

AI compound design

The platform is able to design high-quality derivatives or analogues based on known active compounds, apply deep learning to generate models, match properties and conduct synthesis feasibility analysis, and quickly find high-quality lead compounds.

AI virtual screening

The platform performs virtual screening on a given target. Using deep learning images, sequence recognition algorithms and millions of manually labelled and cleaned active compound data for modelling, it quickly screens the most potential compounds and finds their location in order to obtain the best results on the predicted performance of multiple industry standard datasets.

AI property prediction

The platform utilises graph network models and high-quality ADMET data for QSAR modelling; it deploys a self-attention mechanism to perform weighted labelling of key nature-related fragments for targeted optimisation.

It is an industry-leading technology, with results published at the top AI conference, IJCAI, and in the top journal of computational chemistry, JCIM; in an internal test conducted by the Schrödinger company, it was able to outperform the industry leading software ADMET predictor.

AI synthesis path design

Here the platform designs a synthetic route for a given compound. Using deep learning to generate models, it swiftly recommends the lowest cost and highest success rate synthesis path.

The Aladdin Drug Discovery Platform has already achieved a number of significant discoveries:

  • Seven new mitochondrial autophagy inducers with a screening hit rate of 44%, in research conducted in co-operation with the NO-Age. Two of these inducers have already passed the stage of animal experiments, with the results expected be published in Nature Biotech;
  • 12 new anti-osteoporosis lead compounds were found with a 58% screening hit rate, of which two compounds reached tens of nanomolar activities – this discovery was undertaken in co-operation with Institute of Microbiology, the Chinese Academy of Sciences; and
  • 15 new kinase inhibitors were discovered with a 65% screening hit rate, co-operating with Legion Pharma.


Wade Menpes-Smith, CEO of Aladdin healthcare Technologies, comments: “Aladdin can now officially provide valuable solutions to the pharmaceutical industry via our proprietary AI Drug Discovery platform that can accelerate the drug development process, covering the processes of virtual screening, hit to lead and lead optimisation.”

Aladdin is also working on several promising pipelines for the small molecule drug discovery in the early stage. Small molecule drugs are relatively simple chemical compounds. These compounds have a small molecular mass; and their biggest advantage is their oral bioavailability. Already in May this year, Aladdin achieved two milestones in drug development against age-related diseases and COVID-19, by discovering new optimised compounds with its AIDD platform. The successful release of Aladdin’s proprietary platform confirms once again the capacity of the company to achieve groundbreaking milestones in the field of drug development. This opens the door to many relevant partnerships in the pharmaceutical industry.

Wade Menpes-Smith
Aladdin Healthcare
Technologies Ltd

This article is from issue 15 of Health Europa. Click here to get your free subscription today.

Subscribe to our newsletter


Please enter your comment!
Please enter your name here