Pharmaceutical giants are competing: from AI to quantum computing, the fast track of drug research and development is ready to go

In 1956, John McCarthy proposed the concept of artificial intelligence (AI) at the Dartmouth conference, and the field of AI came into being.

AI-enabled drug development

The development of a new drug is a very long process.

Machine learning uses the rapid processing function of big data to make many analysis processes in drug development more efficient, which may shorten several years of research and development work and save hundreds of millions of dollars in investment.

Currently, AI has been successfully applied in the 4 main stages of drug development:

Stage 1: Determine the target of intervention The first step in drug development is to understand the biological origin (pathway) of the disease and its resistance mechanism, and then determine a good target (usually protein) for disease treatment.

Combining machine learning with multimodal data sets and almost unlimited computing power not only allows researchers to "reconstruct the underlying mechanism of disease", but also makes it easier to analyze all available data, and even learn to automatically identify accurate target proteins .

Representative company: BenevolentAI, Standigm

Phase 2: After discovering the drug candidate and determining the target, the drug developer needs to screen out a compound that can interact with it in a desired way from thousands or even millions of potential compounds, and also need to avoid this compound pair Non-target sites produce side effects.

The currently used screening software has low accuracy and may cause false positive problems, and the process of narrowing the screening range to the best candidate drug is time-consuming.

Machine learning can learn to predict the applicability of molecules based on structural fingerprints and molecular descriptors.

Representative companies: Exscientia, Insilico Medicine

Phase 3: Speeding up clinical trials Clinical trials are the key to the drug development stage.

Machine learning can speed up the design of clinical trials by automatically identifying suitable candidate patients and ensuring the correct distribution of trial participants.

Representative company: IBM Watson, Deep6 AI

Stage 4: Looking for biomarkers for the diagnosis of the disease.

AI can automate a large part of manual work.

Representative companies: ReviveMed, Berg Health

AI can also do: drug reuse, analysis and research literature, publications, patents, etc.

The giants compete

At present, there are more than 200 companies developing AI-assisted drug R&D around the world, and large pharmaceutical companies are also actively deploying through acquisitions and cooperation.

Table 1 Part of AI cooperation of pharmaceutical giants

Data source: official website of each company

In addition, some large technology companies in the Internet and other fields have used their technological advantages to join the pharmaceutical field.

Table 2 Large domestic technology companies that have entered AI pharmacy

Data source: official website of each company

Quantum computing boom is rising

AI development and competition are so heated here, and the research and development of quantum computing-enabling drugs in Nashan is showing signs of development.

Quantum computing relies on quantum computers, which can achieve tasks that traditional computers cannot accomplish.
For example, China's "Nine Chapters" quantum computer can complete the calculations that the world's most powerful supercomputer takes 600 million years to complete in only 200 seconds.

The pharmaceutical giants have already smelled the value.
On January 11, Boehringer Ingelheim announced that it has signed a cooperation agreement with Google Quantum AI to jointly research and implement quantum computing in drug development, especially the application of molecular dynamics simulation.

On January 28, Cambridge Quantum Computing (Cambridge Quantum Computing) announced that it has reached a cooperation with Roche on the design and implementation of the NISQ (noisy-intermediate-scale-quantum) algorithm in early drug discovery and development.
The applied disease field will be Altz In September 2020, Roche announced the failure of the semorinemab jointly developed with AC Immune in the phase II clinical trial.
The choice of quantum algorithm is another new attempt by Roche in the field of Alzheimer's disease.

summary

Whether it is the turn of a large Internet company or the establishment of a new company, each of them tackles one or more aspects of the traditional pharmaceutical process.
Through technological development and transaction cooperation, they will bring development and progress in this field.

How fast drug development can be in the future can only be left to prove.

Reference materials:

1# Artificial Intelligence in Medicine (Source: datarevenue)

2# Eight ways machine learning is assisting medicine (Source: Nature Medicine)

3# How artificial intelligence is changing drug discovery (Source: Nature)

4# Quantum Computing: Boehringer Ingelheim and Google Partner for Pharma R&D (Source: Boehringer Ingelheim official website)

5# Cambridge Quantum to Develop Quantum Algorithms with Roche for Drug Discovery & Development (Source: Cambridge Quantum Computing official website)

6# The company's official website information