9 AI-based initiatives catalyzing immunotherapy in 2018

9 AI-based initiatives catalyzing immunotherapy in 2018

By Tanima Bose

Artificial intelligence (AI) has been dynamically changing the field of drug discovery in several areas, including immunotherapy, oncotherapy, neurological disease, and many more. Described here is an ensemble of the initiatives in place to advance the field of immunotherapy research.   

1) Pfizer-IBM collaboration:

On December 1, 2016, Pfizer and IBM Watson Health announced a collaboration to deploy IBM Watson’s AI platform for drug discovery to accelerate immuno-oncology research.

IBM Watson Health and the Watson Health Cloud Platform were launched by IBM in April 2015; this was the first commercially available platform with cognitive computing capability. The efficiency of this platform is 10,000 times greater than that of a human being. For example, IBM states that an individual can (realistically) read 200 to 300 papers in a given year, whereas Watson can screen 25 million full abstracts, 1 million full papers, and 4 million patents in the same time period. Pfizer hopes to use IBM Watson’s abilities to find hidden connections within the published literature that could lead to novel combination cancer immunotherapies.

2) Evotec investment in Exscientia:

With the ongoing successful collaboration between German-based biotech company Evotec and UK-based AI company Exscientia since early 2016, Evotec announced on September 28, 2017, a €15 million investment in Exscientia to acquire its minority stake. These two companies are aiming to generate new small molecule immuno-oncology drugs using their automated design platform.

For rapid designing of the drug, Evotec plans to combine the AI platform of Exscientia and its own medicinal chemistry platform. Exscientia started its journey in the year of 2013 and collaborated with a number of Pharmaceutical companies like Janssen R&D (2013), Sunovion (2014), Sumitomo Dainippon (2015), and GSK (2017).

3) Microsoft-Adaptive collaboration:

This collaboration was announced on January the 4th of this year with the goal of decoding the immune system to diagnose disease. The technique they will utilize is to collect blood samples, carry out immunosequencing to learn the diagnostic information stored in each immune cell, and then generate an immune map by using machine learning. This immune map will hope to contain information matching trillions of T cells and the diseases they recognize, and this map will be used by doctors and researchers to improve disease diagnosis.

They describe this system as an ‘X-ray of the immune system’ exploring the immunological memory of an individual, and the information on the memory phenotype will help to understand their infection history and exposure to microbes. This new project, named  ‘Decoding the Human Immune System: A Closer Look at a Landmark Partnership,’ is a cornerstone of Microsoft’s Healthcare NExT initiative.

4) Johns Hopkins’ ImmunoMap:

This platform was developed entirely by academic scientists from the Johns Hopkins University. The study was published as a research paper in Cancer Immunology Research on December 20, 2017. The team at Johns Hopkins created a digital map from sequences of T-cells and exposed it to a lab-grown virus.

Using an unsupervised learning algorithm, the team was able to convert the T-cell receptor sequencing data into numeric distances based on similarities in the receptor sequences and cluster them by functional specificity. For example, if two receptor sequences were similar, the computer assigned a short distance rank between the two sequences. If the sequences were different, they received a longer distance rank. Once the thousands of sequences were converted into these “distance” metrics, the computer system’s AI algorithms looked for patterns among the receptors. Their final goal is to find out a homology in T-cells that may target the same antigen.

5) Evince Bioscience’s Autonomous Virtual Organism:

Evince Bioscience is exploiting machine learning to give an innovative facet to drug discovery. Their platform is called Autonomous Virtual Organism (AVO), which hopes to predict the activity, toxicity, and drug-like properties (ADME) in a virtual environment. They initiate their process with the drug-like molecules predicted by AVO and modify it more with the knowledge of medicinal chemistry. Their final goal is to accelerate the process of lead optimization with an accurate prediction by the AVO system.

Their first candidate to test is an orally administered TLR7 drug for cancer immunotherapy. They chose this target because they believe that stimulation of the innate immune system by the stimulation of interferon genes (STING) is one of the recent priorities of immuno-oncology drug discovery.

6) ITUS Corporation’s Cchek:

ITUS corporation, a company focused on designing cancer drugs, has developed a platform called Cchek™, which employs immune profiling by flow cytometry and AI to detect solid tumors from blood samples. The President and CEO of ITUS, Dr. Amit Kumar, is quite hopeful and mentioned in the keynote symposium on February 7, 2018 in San Jose, California:

“The idea of finding cancer in a tube of blood is a very intriguing concept that has generated tremendous interest from the scientific and investment community.”

7) Insilico Medicine:

This Maryland-based start-up uses deep learning, big data, and genomics to design new groups of in silico drugs. Their pipeline drugs target several fields involving cancer immunology as well.

Researchers from Johns Hopkins medical school have recently determined in a study, with the collaboration of Insilico Medicine, that the failure of most of immune checkpoint inhibitors is the production of a transforming growth factor (TGF) by most of the tumor cells. Using AI software, the researchers found out that the TGFβ pathway activation in various cancers is highly correlated with FOXP3, the signature of regulatory T cells (Tregs). Tumors are frequently infiltrated by Tregs, and this is the prominent reason for the poor outcome in multiple cancer types.

8) Texas Advanced Computing Center (TACC) and VDJ server:

Developed by bioinformaticians, immunologists, and computational experts from UT Southwestern Medical Center, J. Craig Venter Institute, Yale University, and TACC, the VDJ server allows researchers to compare discrete, disparate sets of data from a wide range of people in an entire community. Scientists can analyze high-throughput immune repertoire sequencing data over the web. The VDJ server can rightly be used to manipulate immuno-oncology therapy. One of the examples include strategic killing of cancer cells after targeting with genetically engineered T cells.

9) CytoReason:

This is the one of the Israel-based companies pioneering in the AI field who has based its research entirely on the immune system. The company was born depending on its’ 10 years research in the Stanford University and Technion. A recent paper in Nature Biotechnology has investigated the largest analysis of immune cell signaling research which horizoned nearly 3,000 previously unlisted cellular interactions, which also helped rewrite the reference book for immune-focused intercellular communications and disease relationships. CytoReason used this data to predict 335 novel cell-cytokine interactions, providing new clues for drug development.

Limitations and conclusion:

Artificial intelligence has been trying to embrace the field of immunology with its efficient hands. But, too much dependence on any technology can be dangerous in the long run. As Director of Evolution Bioscience, Dr. Frank Rinaldi, has rightfully pointed out:

“It is important, however, not to overstate the potential. AI approaches to drug discovery rely heavily upon accessible high-quality data.”

The ideal way to move forward is to find a strategy that will combine new technology with traditional methods.

If successful, artificial intelligence will help to resolve the drug discovery scientific questions in the field of both oncotherapy and infectious diseases. Another important thing to point out is that the older researchers with training in conventional methods have to be updated themselves with the newer techniques, and this will lead to the holistic improvement in the field. In conclusion, we envision artificial intelligence will bring a brighter future to the field of immunotherapy.  

If you have any questions or would like to know if we can help you with your innovation challenge, please contact our Life Sciences lead, Jeremy Schmerer at jschmerer@prescouter.com or our Strategic Accounts Manager, Linda Cohen at lcohen@prescouter.com.

More of Our Insights & Work