Most Innovative R&D Approaches of AI in Biopharma
The industry of AI in Biopharma continues growing up after a long period of skepsis, which is reflected in a substantial increase in the volume of investments and the number of joint ventures in 2019 compared to 2018 and earlier years. The difference between Pharma and Biopharma fields is that biopharma medicines and drug products are manufactured in living organisms like bacteria, yeast and mammalian cells. The prefix “bio” refers to how drugs are produced. Biopharma is the subset of drugs produced by biological methods. Pharmaceutical drugs cover biological means as well as chemical synthesis.
The Biopharma industry’s growth dynamics is largely influenced by the more active participation of largest pharmaceutical corporations in the AI-related investment and research collaborations. Despite some Pharma corporations still being critical about AI applications, the number of researches, scientific publications in the field of AI in Biopharma and research collaborations between pharma companies and AI-expertise vendors is rapidly increasing.
Nevertheless research in AI is facing challenges today, but the demand for the ML/AI technologies, as well as for ML/AI talent, is growing in pharmaceutical and healthcare industries and driving the formation of a new interdisciplinary field — data-driven drug discovery/healthcare. The overall success of all the companies in the industry depends strongly on the presence of highly skilled interdisciplinary leaders, able to innovate, organize and guide in this direction. It will be crucial to hire top AI experts, especially for Big Pharma companies that are fighting to survive.
Trending and most innovative R&D approaches of top AI in Biopharma companies include application of:
Natural Language Processing
Cognitive Reasoning Technologies
20 R&D Approaches of AI in Biopharma
How Abbvie Uses AI in R&D?
AbbVie is a global, research-based biopharmaceutical company formed in 2013 following separation from Abbott. The company's mission is to use its expertise, dedicated people and unique approach to innovation to develop and market advanced therapies that address some of the world's most complex and serious diseases.
There is not much information about use of artificial intelligence in drug discovery by AbbVie. But it does have a confidential project listed with Atomwise. Also, in September 2016, together with its partner AiCure, AbbVie announced how its AI-based patient monitoring platform improved adherence in an AbbVie phase 2 schizophrenia trial.
Abbvie and Mission Therapeutics collaboration is aimed on developing DUB inhibitors that promise to treat two currently incurable conditions, Parkinson’s and Alzheimer’s diseases. With over 50 million Americans struggling Alzheimer’s and dementia, this AI partnership will bring treatment closer and hope for many.
Abbvie cooperates with another AI-specialized company AiCure, a clinically-validated artificial intelligence company that visually confirms medication ingestion on smartphones, announced that study results presented today during the International Society for CNS Clinical Trials and Methodology (ISCTM) Scientific Sessions confirm that use of the AiCure Platform significantly increases medication adherence in patients with schizophrenia, as measured by drug concentration levels.
The sub-study was part of a larger Abbvie, Phase 2, multicenter, randomized, double-blind, placebo-controlled, dose-ranging, parallel-group, study in nonsmoking subjects with schizophrenia who were clinically stable. Subjects were enrolled and randomized to placebo or ABT-126. The AiCure platform was introduced into 10 of 31 US sites; subjects were monitored either by AiCure or by modified Directly Observed Therapy (mDOT) at least 3 times per week. In addition, adherence was measured by review of returned study drug blister and scheduled pharmacokinetic sampling.
Results: cumulative adherence, measured by study drug concentrations above the LLOQ (minimum required therapeutic level), were higher through 24 weeks for subjects monitored using the AiCure platform (89.7%) compared with subjects monitored using mDOT (71.9%). This research adds to the growing body of scientific evidence showing the advantages of using AI to increase statistical power and reduce sample size in clinical trials, thereby decreasing costs and accelerating drug development.
Subjects of the investigation were enrolled and randomized to placebo or ABT-126
Subjects were dosed once daily (QD) 25, 50, or 75mg ABT-126 or matching placebo, as 3 capsules in the morning for 24 weeks
Subjects were monitored either by AiCure or by modified Directly Observed Therapy (mDOT) at least 3 times per week
Adherence was measured by review of returned study drug blister and scheduled pharmacokinetic sampling
How AstraZeneca Uses AI in R&D?
AstraZeneca is a global, science-led biopharmaceutical company, it`s innovative medicines are used by millions of patients worldwide. Jim Weatherall, Vice President in AstraZeneca, said that data science and AI has the potential to transform the way they develop new medicines – turning yesterday’s science fiction into today’s reality with the aim of enabling the translation of innovative science into life-changing medicines.
At AstraZeneca, they are using AI to combine information from multiple sources researchers hope to draw more accurate conclusions than if they analysed science literature by hand. AI also has the potential to find patterns in these graphs revealing previously unexplored hypotheses.
AstraZeneca focuses on the discovery, development and commercialisation of prescription medicines, primarily for the treatment of diseases in three therapy areas - Oncology, Cardiovascular, Renal & Metabolism and Respiratory. The company has turned to AI to cut development costs by improving the efficiency of repetitive tasks and engendering better-informed decision.
The knowledge graphs are used to give their scientists the information they need about genes, proteins, diseases and compounds and how they relate to each other. Using AI to combine information from multiple sources they hope to draw more accurate conclusions than if researchers analysed science literature by hand. AI also has the potential to find patterns in these graphs revealing previously unexplored hypotheses.
Discovering a potential drug molecule requires several years of detailed scientific research. AI is enabling us to rapidly generate novel ideas for molecules to make and rank these ideas using predictions based on large data sets available to us. Having identified promising molecules, the next step is to synthesise the molecules in the laboratory. AI is starting to help here too – the science of synthesis prediction is rapidly evolving and scientist will soon be able to use AI to help deduce the best way to make a molecule in the shortest time.
AI systems are trained to assist pathologists in analysing samples accurately and more effortlessly. This has the potential to cut analysis time by over 30%. For one of their AI systems, they implemented an approach inspired from how some self-driving cars understand their environment. They trained the AI system to score tumour cells and immune cells for a biomarker, called PD-L1, which has potential to help inform immunotherapy-based treatment decisions for bladder cancer.
Researchers can ask key questions to help identify and prioritise drug targets
How AMGEN Uses AI in R&D?
Amgen is one of the world’s leading biotechnology companies. Amgen is committed to unlocking the potential of biology for patients suffering from serious illnesses by discovering, developing, manufacturing and delivering innovative human therapeutics. This approach begins by using tools like advanced human genetics to unravel the complexities of disease and understand the fundamentals of human biology.
AI is but one of a series of emerging digital capabilities Amgen is advancing to improve how they do a whole host of activities across the company — from drug discovery and patient identification to optimized interactions with physicians. Other technologies that company is leveraging include digital automation, natural language processing, advanced analytics.
Amgen is piloting a process using AI that has the potential to greatly enhance its ability to trend and find patterns in manufacturing deviations and to prevent their recurrence. The AI tool will replace a manual, labor-intensive process with one that can look across large data sets and find correlations between obscure signals and events which the previous system could have missed.
1. While large company manufacturing, purifying, and packaging biotech drugs, a huge amount of diverse data is generated, not all of which is digitized. The focus of Amgen is the application of data science specifically in quality operations, using a data science process:Quality data sciences creates solutions that unlock and leverage data. These solutions will efficiently provide insights and intelligence for the Quality Operation. This involves: 1) Ensuring data access; 2) Application of appropriate analysis methods to unlock information; 3) Meaningful visualisation
2. Amgen have created a project team to look for a system algorithm that could replicate and perhaps improve upon the manual process. The goal was to think big but start small and build a product that could be deployed across the manufacturing network. Using an agile development approach and natural language processing (NLP) tools, the team developed a consistent algorithm that was able to reasonably replicate the manual process.
NLP is described as an AI technology that turns text into numbers, which can be read by a computer and used to identify similar records. Each record has a series of numbers associated with it that can be analyzed to create similarity scores. The records can then be clustered together. Those clusters can then be given to an subject matter expert, who can decide if there is trending and if action should be taken. Feedback can then be given to the algorithm, which can be adjusted.
Get the data
Explore the data
Model the data
Visualise the results
Ask an interesting question
How BenevolentAI Uses AI in R&D?
BenevolentAI is the global leader in the application of AI for scientific innovation. The company's aim is to accelerate the journey from inventive ideas to medicines for patients by developing AI to generate new treatments for some of the world’s 8,000 untreated diseases. BenevolentAI integrates AI technologies at every step of the drug discovery process: from early discovery to late stage clinical development.
The company has developed the Benevolent Platform™ - a leading computational and experimental discovery platform that allows our scientists to find new ways to treat disease and personalise medicines to patients. The Benevolent Platform™ focuses on three key areas: Target Identification, Molecular Design and Precision Medicine.
The way AI is used:
to force the discovery of drug patterns;
to collect more diverse data;
to identificate specific target;
in molecular design;
in patient stratification.
The Benevolent Platform™ of computational and experimental technologies and processes, draws on vast quantities of mined and inferred biomedical data and is built and used by their world-class scientists, researchers, and technologists, working side-by-side, to improve and accelerate every steps of the drug discovery process.
BenevolentAI uses AI to mine and analyse biomedical information, from clinical trials data to academic papers. The company's approach:
Hypothesis generation and target identification
AI-Augmented molecular design
Benevolent Knowledge Graph including algorithmically derived knowledge
Patient stratificationvfor better clinical trials
2. BenevolentAI has spent the last five years developing a knowledge pipeline that pulls data from various structured and unstructured biomedical data sources and curates and standardizes this knowledge via a data fabric.
This is fed into our proprietary knowledge graph which extracts and contextualises the relevant information.
The knowledge graph is made up of a vast number of contextualised, machine curated relationships between diseases, genes, drugs and with over 20 types of biomedical entities.
3. Relation inference AI models help to predict potential non-obvious disease targets that may be overlooked. Their specific expression based models help to identify proteins, genes that express differently in a disease and healthy cell.
4. By leveraging advanced AI, the EvoChem product designs de novo compounds based on multiparametric optimisations with a scoring function that factors in all the properties the company is seeking to optimise for that molecule.
5. Company applies ML models to identify patient groups by the molecular signature of their disease and design, allowing to run faster clinical trials.
How Boehringer Ingelheim
Uses AI in R&D?
Boehringer Ingelheim is one of the world's largest pharmaceutical companies, and the largest private one. The company's key areas of interest are: respiratory diseases, metabolism, immunology, oncology and diseases of the central nervous system.
The focus of the company’s AI-related activity is in doing so is on diseases for which no satisfactory treatment option exists to date. The company therefore concentrates on developing innovative therapies that can extend patients’ lives. In animal health, Boehringer Ingelheim stands for advanced prevention.
The way AI is used:
to boost the efficiency speed;
to reduce the time needed to discover a new drug;
to improve the quality of discovered drugs and molecules.
Boehringer Ingelheim has partnered with UK-based AI tech company Bactevo to speed up its drug discovery efforts. In this collaboration, Boehringer will leverage Bactevo’s AI-powered platform – Totally Integrated Medicines Engine platform (TIME) – to boost the efficiency, speed, and quality of drug discovery from small molecule lead compounds. As a result they obtain the reduction of time to take drugs to market for treatment of conditions caused by defects in mitochondrial function. It essentially brings together the powerful drug research experience at Boehringer and state of the art TIME drug discovery platform to discover new medicines for ALS, Parkinson’s disease and Alzheimer’s disease.
With the founding of BI X as independent subsidiary Boehringer Ingelheim will focus on breakthrough innovative digital solutions in healthcare from idea to pilot. The start-up will work closely together with all three business units of the company - Human Pharma, Animal Health and Biopharmaceuticals. It will provide a platform for collaborating with specialists in the field of data science, agile software development and user experience design.
BI X will develop prototypes for new products and solutions and test them together with the company's business units in pilot phases. The business units will then use the successfully developed new products and solutions themselves and bring them to the market. This approach is to ensure that knowledge and experience accrued at BI X are being quickly integrated into the digital lab's parent house.
Analytical algorithms in audio tools can lead to advances in the earlier diagnosis of diseases in humans and animals. SoundTalks, an audio monitoring system for the early detection of respiratory diseases, is currently being tested in livestock farming. In humans as well, faster treatment can slow or even halt the progression of a disease – which is essential, particularly for central nervous system disorders. Intelligent speech recognition software – via smartphone, for example – will be able to analyze speech patterns, recognize risks, and thereby contribute to a reliable diagnosis and effective therapy.
How Celgene Uses AI in R&D?
Their vision as a company is to build a major biopharmaceutical corporation while focusing on the discovery, the development, and the commercialization of products for the treatment of cancer and other severe, immune, inflammatory conditions. There are more than 300 clinical trials at medical centers using compounds from Celgene. The company is transforming pharmacovigilance (PV) to drive the new era of patient safety.
Pharmacovigilance detects, assesses, and prevents adverse events (AEs) and other drug-related problems by collecting, evaluating, and acting upon AEs. The value of using AI methodologies in PV is compelling; however, as PV is highly regulated, acceptability will require assurances of quality, consistency, and standardization.
The way AI is used:
to speed up the discovery of drug candidates for for cancer and autoimmune diseases;
to identify and standardize PV knowledge elements;
to develop, review and validate cognitive services;
to increase operational efficiency, consistency, quality of data collection, and signal detection.
Celgene force boost AI implementation in drug discovery particularly machine learning and deep learning. Machine learning involves computing techniques that analyze vast amounts of data to find understanding that might be too abstract and time consuming for humans. Deep learning takes that even further, using code that attempts to mimic the brain’s ability to recognize patterns in unstructured data.
In the past, researchers relied on imperfect image-processing algorithms to analyze cancer cells, and then they corrected them by hand. With tens of thousands of cells, this required a huge expenditure of time and effort. But using deep learning, images can be processed almost instantaneously with much better results. For these analyses use:
Amazon SageMaker machine-learning platform
Virtually analyzing the biological impact of a potential drug
*without putting live patients (or even lab rats) at risk
Apache MXNet deep-learning framework
2. Pharmaceutical research revolves heavily around exceedingly complex algorithms to predict how certain compounds will interact with the human body. To this end, Celgene uses high-performance Amazon EC2 P3 instances powered by NVIDIA Tesla V100 Tensor Core GPUs (graphics processing units) to process the complexity. These NVIDIA GPUs have thousands of cores that accelerate the training of machine-learning models (which can, for instance, test the effectiveness of a drug at faster and more accurate rates). The results have been game changing: A model that once took two months to train can now be trained in four hours.
3. AI is used to identify areas across the pharmacovigilance (PV) value chain that can be augmented by cognitive service solutions using the methodologies of contextual analysis and cognitive load theory. It will also provide a framework of how to validate these PV cognitive services leveraging the acceptable quality limit approach.
How GlaxoSmithKline Uses AI in R&D?
GlaxoSmithKline has 3 global businesses - Pharmaceuticals, Vaccines, Consumer Healthcare - that research, develop and manufacture innovative pharmaceutical medicines, vaccines and consumer healthcare products. Their R&D approach focuses on science related to the immune system, use of genetics and advanced technologies.
GlaxoSmithKline has many deals with different companies such as Exscientia, Insilico Medicine, Insilico Biotechnology to use new computer modelling systems to bring differentiated, high-quality and needed healthcare products.
The way AI is used:
to improve the discovery of drugs, biomarkers, and new vaccines
to interpret and understand genetics and genomic data;
to understand the effect of interventions on diseases.
The goal of GSK is to achieve a sustainable flow of meaningful new treatments, utilising modalities such as small molecules, antibodies, antibody drug conjugates and cells, either alone or in combination. Their research focuses on science related to the immune system and human genetics, while leveraging advanced technologies including functional genomics, AI and machine learning.
According to the GSK and Exscientia collaboration, AI enabled platform will be applied and combined with the expertise of GSK, in order to discover novel and selective small molecules for up to 10 disease-related targets, nominated by GSK across multiple therapeutic areas.
GSK evaluates Insilico Medicine technology in the identification of novel biological targets and pathways of interest to GSK to enhance its drug discovery process.
GSK uses Insilico' Biotechnology technology platform to analyse predictive simulations of cell responses. As a result an almost unlimited number of scenarios will be generated and utilized to significantly reduce the number of experiments necessary during bioprocess development. This has the potential to cut down the time taken to research Vaccine manufacturing processes and could therefore speed up the time-to-market for candidate Vaccines in GSK's research portfolio.
Insilico's technology is built around Insilico Cells™:
Genome-based mechanistic network models of common host organisms
Models are refined and calibrated
Models become individualized representation of the specific cells in silico
Simulation of cells behavior in a bioprocess
How Gilead Uses AI in R&D?
Gilead is focused on developing and delivering medications that advance the treatment of life-threatening diseases. The commercial success of their products provides them with the resources to generate new clinical data defining their profiles and supports their development of new therapeutic advancements. As they bring new products into clinical development, their goal remains the same – to discover, develop and commercialize therapeutics that advance patient care.
In spring 2019 Gilead and insitro announced that the companies have entered into a strategic collaboration to discover and develop therapies for patients with nonalcoholic steatohepatitis (NASH).
The way AI is used:
to сreate disease model;
to discover new targets that influence disease progression and regression.
According to John McHutchison, the company’s Head of Research and Development, Gilead is committed to researching and developing treatments for patients living with NASH, particularly those with advanced fibrosis who have the greatest unmet need. Gilead is able to utilize deep learning to explore the scientific underpinnings of the biology and clinical spectrum of NASH, with the goal of accelerating the development of highly effective treatment options for patients with this disease.
The startup's insitro Human platform combines AI with human genetics and genomic data to provide insights into the disease's makeup and progression, propose forms of treatment and predict patient responses to those therapies. With that information, Gilead is planning to chemically develop up to five of the proposed treatments for NASH.
By generating high-throughput, functional genomic data sets that align with patient data, and interpreting those data via novel machine learning methods, insitro builds predictive models that can accelerate target selection and the design of effective therapeutics. The company is building a high-throughput bio-data factory based on state-of-the-art technologies from bioengineering, allowing the creation of large data sets that enable cutting edge machine learning methods to be brought to bear on key bottlenecks in drug development.
Gilead datasets from patients and preclinical trials
Insitro IHS platform
Data interpretation with machine learning methods
Design of new drugs & therapeutics
3. The insitro Human (ISH) platform applies machine learning, human genetics and functional genomics to generate and optimize unique in vitro models and drive therapeutic discovery and development. The ISH platform provides insights into disease progression, suggest candidate targets, and predict patient responses to potential therapeutic interventions. Gilead can advance up to five targets identified through this collaboration and is responsible for chemistry and development against these targets.
How Insilico Medicine Uses AI in R&D?
Insitro IHS platform
Insilico Medicine, Inc. is a bioinformatics company located at the Emerging Technology Centers at the Johns Hopkins University Eastern campus in Baltimore. It utilizes advances in genomics, big data analysis and deep learning for in silico drug discovery and drug repurposing for age-related diseases. The company pursues internal drug discovery programs and geroprotector discovery and provides services to pharmaceutical companies.
Combining genomics, big data analysis, and deep learning, the company has been using artificial intelligence algorithms to potentially discover the next world-changing drug.
The way AI is used:
to cheaper and faster discover of drug molecules;
to imagine new molecules with drug-like properties;
to find new drug candidates, biological targets and molecules;
to validate the targets, using novel chemistry.
Juvenescence AI Limited
Next-generation AI developed by Insilico Medicine can be used to validate, assess and improve the quality of biological samples as well as learn using large volumes of heterogeneous data without human intervention. Multiple new methodologies including the feature importance, deep feature selection and deep pathway analysis among the others can provide the biologically-relevant interpretation of the inner workings of the AI systems.
1. Since 2016, Insilico Medicine researchers have been working to get GANs (Generative Adversarial Networks consisting of two distinct neural networks) to “imagine” new molecules with drug-like properties. In 2017, they combined this with another type of groundbreaking A.I. in the form of Reinforcement Learning. Reinforcement Learning is built around the notion of A.I. agents which use trial-and-error to maximize some kind of reward.
in vitro assays
in vivo assays
2. Insilico Medicine has developed GENTRL (Generative Tensorial Reinforcement Learning), a new artificial intelligence system for drug discovery that dramatically accelerates the process from years to days (from 3 years to 21 days before first synthesis and trials). In the industry’s first successful experimental validation of such AI technology for drug discovery in cells and animals, Insilico successfully tested the technology by creating a series of entirely new molecules capable of combating disorders like fibrosis.
3. The system bucks the standard brute-force approach for AI drug development, which involves screening millions of potential molecular structures looking for a viable fit, in favor of a creative AI algorithm that can imagine potential protein structures based on existing research and certain preprogrammed design criteria. Insilico's system initially produced 30,000 possible designs, which the research team whittled down to six that were synthesized in the lab, with one design eventually tested on mice to promising results.
How Nuritas Uses AI in R&D?
Nuritas is revolutionising the discovery of novel, natural and scientifically proven active ingredients that can manage and improve human health. The company’s disruptive computational approach to discovery uses artificial intelligence and genomics to, for the first time ever, rapidly and efficiently predict and then provide access to the most health-benefiting components hidden within food, called bioactive peptides.
Their bioactive peptides provide patented innovative solutions to companies needing new therapeutic options to deal with significant unmet medical needs. The Bioactive Peptides they discover have the potential to offer new and innovative treatments for many of the illnesses that are becoming more prevalent as the world population continues to expand and age.
The way AI is used:
to target, predict and unlock novel bioactive peptides;
to deliver highly specific, efficient and life-changing health solutions;
At Nuritas, they recognise the vast untapped potential that exists in naturally occurring Bioactive Peptides. Their unique discovery platform targets, predicts and unlocks these natural ingredients to provide new solutions and opportunities for their partners across a wide range of application areas. Together with their partners, they are changing the lives of billions of people worldwide.
1. They begin the discovery process by precisely defining the health condition and targets they wish to modulate. They then use proprietary search tools to identify the characteristics specific to their area of focus. The most up-to-date academic and scientific knowledge is used to maximise the efficiency and effectiveness of prediction algorithms.
Naturally Sourced Ingredients
High Speed Prediction
Defining of targets to modulate
Identifying of key characteristics of the focus area
Predicting of bioactive peptides’ effect using deep learning algorithms
Unlocking peptides from food source
Cost Effective Innovation
2. Having begun the discovery process as above, Nuritas takes advantage of multiple proprietary AI algorithms, including deep learning. Using these, they are now uniquely able to predict which novel food-derived bioactive peptides deliver the pre-determined effect that they are seeking. This cuts out many thousands of hours of trial and error.
3. After targeting and predicting high potential Bioactive Peptides, Nuritas unlocks them from within the food source for their pre-defined therapeutic use.
4. Their library of plant and animal derived Bioactive Peptides have gone through hundreds of millions of years of selective evolution to become the most potent repairers, healers and protectors.
How Novo Nordisk Uses AI in R&D?
Novo Nordisk is a global healthcare company with more than 95 years of innovation and leadership in diabetes care. This heritage has given us experience and capabilities that also enable us to help people defeat other serious chronic diseases: haemophilia, growth disorders and obesity. At Novo Nordisk, they are driving change to defeat diabetes and other serious chronic diseases.
Novo Nordisk cooperates with e-Therapeutics to use its AI-based drug discovery technology to find new therapies for type 2 diabetes. e-Therapeutics uses a suite of powerful computational tools to augment and interrogate the vast amount of biological information currently available in both public and private databases.
The way AI is used:
to identify novel intervention strategies;
to find new biological pathways and compounds;
tease out previously unknown disease processes and pathways;
to form the basis for new therapies.
The company said this approach more realistically reflects the true complexity of disease with its multiple and often interconnected cellular pathways. Novo Nordisk already has a research centre in Oxford, where visiting researchers are working with Oxford University academics to advance development of therapies for type 2 diabetes.
Using techniques such as machine learning and state of the art data analysis, e-Therapeutics creates and analyses network models of disease to identify likely proteins that could be disrupted to treat diseases. Company’s scientists use AI to improve and accelerate active pharmaceutical ingredient (API) and drug product (DP) designs of new peptides and biologics. They have access to advanced state-of-the-art protein modelling software and are applying data science tools and prediction models to a variety of biological and chemical data, including high-throughput analytical data and images.
AI-related research department structure
Modelling of molecules and proteins in order to create better and more stable drugs
Understanding and applying of complex machine learning algorithms
AI-related research department structure
Ensuring that entire working infrastructure is as agile and stable as possible
2. Novo Nordisk is the first pharma partner to sign up to use the GAIN platform, which taps into genome-wide association study (GWAS) data to find mutations in DNA linked to disease traits and – according to the UK company – bridges the gap between genetic susceptibility and disease mechanism. While many gene variants discovered using GWAS studies often don’t map to a plausible biological mechanism, e-Therapeutics says its “network biology” approach can improve the hit rate.
3. Using GAINs, the company will be able to interrogate genomics data from patients with complex, polygenic disease and shed new light on important and novel biological pathways for particular groups of patients.
How Pfizer Uses AI in R&D?
Pfizer is a leading research-based biopharmaceutical company. They apply science and their global resources to deliver innovative therapies that extend and significantly improve lives. The company's products are the results of 1500 scientists overseeing more than 500,000 lab tests and over 36 clinical trials before the first prescription.
Pfizer in late 2016 announced a collaboration that will utilize IBM Watson for Drug Discovery. Pfizer is using IBM’s AI technology on its immuno-oncology research, a strategy of using a body’s immune system to help fight cancer. Based on their research, this appears to be one of the first significant uses of Watson for drug discovery.
The way AI is used:
to uncover new information or insights related to patient needs;
to analyze massive volumes of disparate data sources, including licensed and publicly available data;
to discover new drug targets and alternative drug indications.
Pfizer is among the healthcare companies investing heavily in AI and teamed with IBM Watson to identify better targets for cancer during the discovery phase and Concerto HealthAI to to apply real-world datasets and artificial intelligence techniques to develop new and more precise treatment options for patients with solid tumors and hematologic malignancies.
Pfizer is one of the first organizations worldwide to deploy Watson for Drug Discovery, and the first to customize the cloud-based cognitive tool:
Natural language processing
Cognitive reasoning technologies
to support the identification of new drug targets, combination therapies for study, and patient selection strategies in immuno-oncology
2. Pfizer uses newly launched Watson for Drug Discovery, a cloud-based offering that aims to help life sciences researchers discover new drug targets and alternative drug indications. The average researcher reads between 200 and 300 articles in a given year 2, while Watson for Drug Discovery has ingested 25 million Medline abstracts, more than 1 million full-text medical journal articles, 4 million patents and is regularly updated. Watson for Drug Discovery can be augmented with an organization's private data such as lab reports and can help researchers look across disparate data sets to surface relationships and reveal hidden patterns through dynamic visualizations.
3. AI systems are used in progressive ways to analyze data, to uncover new information or insights related to patient needs.
How Roche Uses AI in R&D?
To intensify it`s on healthcare, Roche divests two businesses: fragrances and flavours, and vitamins and fine chemicals. As a research-driven company committed to innovation, the Group’s Pharmaceuticals and Diagnostics Divisions supply products spanning the healthcare spectrum, from the early detection and prevention of disease to diagnosis and treatment.
Sensors, wearables, IOT, blockchain, high performance compute, Machine Learning and Deep Learning are drivers and enables of digital transformation of Roche`s entire Pharma value chain. AI is expected to have a dramatic impact on medicine that Roche provides.
The way AI is used:
to improve the ability to diagnose disease;
to select the best treatments for individual patients;
to De novo compound design;
to better target selection.
With the advent of more sophisticated digital technologies, personalised healthcare is entering a new phase, expanding from companion diagnostics to a more complex, holistic view of patient health generated from a wide variety of data sources. Combining and standardising data, using AI and algorithms to make sense of it all, enhance the way Roche can develop and bring medicines to patients in a much more targeted fashion.
Examples of transformative digital use cases across different areas, modalities and pipeline phases:
Molecule discovery and optimization
Deep image analysis in Ophthalmology
AI and Wearables
Digital biomarkers in Neuroscience
2. AI could revolutionise the way ophthalmologists diagnose diabetic macular edema (DME), a complication of diabetes that causes a thickening of the retina that can lead to irreversible blindness if left untreated. The best way to prevent DME is through regular eye exams that use a technique called colour fundus photography (CFP) and optical coherence tomography (OCT). The company's researchers use deep learning to teach computers how to estimate macular thickness from CFP images, making DME diagnosis easier, so they gave their computers a large set of CFP and OCT data from participants in two large DME clinical trials to train on.
The deep learning system examined a total of 17,997 CFP images from ~700 patients and compared them with corresponding OCT thickness measurements. Deep learning could even do a reliable job of predicting the actual OCT measurement of the macula’s thickness from a CFP image if it was of sufficient quality.
How Sanofi Uses AI in R&D?
Sanofi is a healthcare company engaged in the research, development, manufacturing, and marketing of innovative therapeutic solutions. It covers areas such as diabetes solutions, human vaccines, innovative drugs, consumer healthcare, etc. Its products includes prescriptions and over-the-counter drugs for thrombosis, cardiovascular disease, diabetes, central nervous system disorders, oncology and internal medicine, vaccines.
Sanofi and Google apply artificial intelligence (AI) across diverse datasets to better forecast sales and inform marketing and supply chain efforts. Using AI will take into account real-time information as well as geographic, logistic and manufacturing constraints to help the accuracy of these complex activities.
The way AI is used:
to provide remote connection between patients and doctors;
to diagnose diseases at early stage;
to decentralize clinical trials;
to improve marketing strategies.
Big data and a better understanding of the human genome are also providing medical professionals with better tools to make faster, more accurate diagnoses and deliver more personalized treatments. Large, multinational databases of clinical data called patient registries can also play a key role in the study of rare diseases. Additionally, Sanofi IT will be modernizing its infrastructure by migrating some existing business applications to Google Cloud Platform (GCP).
1. MouthLab is a single, noninvasive device that measures more than 10 different health indicators in less than a minute. The AI-powered system uses the patient’s mouth and hand to measure in real-time vital health signs typically monitored at the doctor’s office, including respiratory rate, pulse, electrocardiogram, blood oxygen saturation, temperature, blood pressure, and several lung functions. In addition, the device connects to the cloud, so patient data is accessible in real time to physicians and caregivers. By making this data easily available, Sanofi aims to reduce hospitalizations, patient costs and risks.
2. Wavy Assistant delivers continuous real-time heart health monitoring using voice and AI solutions. After a patient’s data is collected and analyzed, Wavy can provide advice tailored to that individual. Our heart health monitoring solution uses a smart home speaker as its main user interface, which allows customers to interact with their heart health easily through a natural conversation instead of a mobile app. If the system detects something is wrong, Wavy instantly sends an emergency signal to designated doctors, friends and family. It can also trigger an immediate alert during emergency situations. Almost all heart attacks and strokes happen at home and most of the damage occurs because the emergency services are called too late.
4. Сomputers that understand humans through text and voice are Sanofi’s solution for healthcare with many applications. Voice technologies are used to detect, e.g. flu—or a general decline in condition—before it gets worse (e.g. pneumonia). A device in the elderly person's home can analyze changes in a person's voice and detect symptoms early on. It helps homecare and home nurses to detect their patients’ illness before they need hospital care.
5. Mentalab combines a wearable patch that can measure electrocardiogram biosignals continuously, with a cloud-based analysis service to diagnose and monitor cardiac and respiratory conditions. The patch can be applied by patients directly, and worn throughout their daily activities, while data is transmitted and analyzed seamlessly. In site-less clinical trials, this solution can increase patient engagement and participation rates.
6. NeuroAdvise is a clinical decision support tool available as a mobile application that helps physicians make better clinical decisions. Our system can archive all demographic and clinical patient-related information in a classified manner. Data is currently stored without patient identity according to time and date. NeuroAdvise algorithms are simulations of a clinician’s mental diagnostic process and most of the important diagnostic factors are included in its comprehensive database. It only takes a few seconds for the user to access the list of differential diagnosis, which are sorted in order of probability and unique for each patient. The system is flexible with unlimited capacity for adding new symptoms, disorders and diagnostic tests.
7. A ring-design cardio tracker, or CART, can provide continuous monitoring of vital signs in real-world clinical trials and can be worn easily in daily life.
How Teva Uses AI in R&D?
Teva is investing in both original biologic medicines and in biosimilars (highly similar versions to specific innovator biologics) to help patients around the world. We are focusing on treatments for the central nervous system, respiratory conditions, and in the field of oncology. Teva uses the help of AI for development of ‘Single device location-algorithm pair’ for optimal treatment of impaired mobility resulting from ageing and chronic disease as well as for digital technology, including body worn sensors.
In cooperation with IBM Research, Teva focuses on two key healthcare areas: the discovery of new treatment options and improving chronic disease management.
The way AI is used:
to improve chronic disease management;
to discover new treatment options;
to repurpose already existing drugs
In cooperation with IBM Research, Teva focuses on two key healthcare areas: the discovery of new treatment options and improving chronic disease management. Teva’s AI projects are based on the IBM Watson Health Cloud, that is a health-data enabled platform-as-a-service which is designed to help healthcare organizations derive individualized insights and obtain a more complete picture of the many factors that can affect people’s health based on machine learning.
1. Teva has chosen the IBM Watson Health Cloud as a preferred global technology platform and managed to build solutions designed to help millions of individuals worldwide with complex and chronic conditions such as asthma, pain, migraine and neurodegenerative diseases. In addition, a joint Teva-IBM Research team will deploy Big Data and machine learning technology to create disease models and advanced therapeutic solutions.
2. Watson is a groundbreaking cognitive computing platform that represents a new era of computing based on its ability to interact in natural language, process vast amounts of Big Data to uncover patterns and insights, and learn from each interaction. The Watson Health Cloud provides an open development platform for physicians, researchers, insurers and companies focused on creating health and wellness solutions.
3. IBM’s Global Business Services works closely with a Teva Analytics team to assess the data and the analytics model requirements for the Real World Evidence e-health solution.
4. By building on the Watson Health Cloud, Teva is in a unique position to put the best information and insights in the hands of physicians, care teams and patients, to empower treatment optimization for individuals and populations across the spectrum of acute and chronic conditions. Watson provides Teva with better insights, real-time feedback and options for clinicians to consider to improve patient care.
How Deep Genomics Uses AI in R&D?
Deep Genomics is using artificial intelligence to build a new universe of life-saving genetic therapies. The company has trained AI algorithms to better understand and search new drug targets in human genomic sequences, specifically those that have been deemed “undruggable,” as well as to design new Antisense Oligonucleotides (ASOs) drugs.
The way AI is used:
Find drug candidates (ASOs) with desirable properties;
To predict molecular phenotypes alterations, such as transcription, splicing, translation and protein binding.
To produce On-target and genome-wide off-target effect data, cell viability data and animal toxicity data.
Deep genomics use its AI Workbench to rapidly discover and develop genetic therapies, and to do so with a increasing success rate.
Genetic Medicines. Deep Genomics’ AI Workbench enables them to efficiently find drugs with desired properties. The company is focussing on the development and marketing of antisense oligonucleotide therapies that target the disrupted genes that cause diseases at the level of RNA or DNA. Deep Genomics is predicting altered molecular phenotypes, such as transcription, splicing, translation and protein binding that may caused genetic diseases.
The Deep Genomics platform is able to produce On-target and genome-wide off-target effect data, cell viability data and animal toxicity data for every compound. They also collect data related to biomarkers. All data is processed using feedback loops.
The Deep Genomics’ research works have appeared in Science, Nature, Nature Genetics, Nature Medicine, Nature Methods, Proceedings of the IEEE, NIPS, Bioinformatics, RECOMB and ISMB.
Project Saturn has proved the utility of Deep Genomics’ AI platform
In Project Saturn, the Deep Genomics’ team was using their platform to evaluate over 69 billion oligonucleotide molecules against 1 million targets in silico, to generate a library of 1000 compounds that were experimentally verified to manipulate cell biology as had intended.
How AI Therapeutics Uses AI in R&D?
AI Therapeutics is an AI-driven company with a unique ability to match drugs according to indications and prosecute through clinical development. AI Therapeutics’ revolutionary approach to drug discovery and development has led to clinical trials of three drugs.
The way AI is used:
to precisely match drugs to diseases;
to identify the best care for individual patients.
Drugs currently in clinical trials target:
B-cell non-Hodgkin lymphoma;
acute myeloid leukemia;
amyotrophic lateral sclerosis;
AI Therapeutics has developed a new Guardian Angel™ artificial intelligence algorithm that has learned to predict new therapies for diseases with unprecedented accuracy. Guardian Angel ™ combines public and proprietary data and is designed to search for drugs for any indication.
How Recursion Pharmaceuticals Uses AI in R&D?
Recursion Pharmaceuticals is a digital biology company, which applies AI for the development of its drug discovery platform and pipeline. The company is Reengineering drug discovery by taking a target-agnostic approach that combines automation, machine learning and the world’s largest biological image dataset with a highly cross-functional team to discover transformative new treatments.The company currently has 4 clinical stage programs and 6 preclinical candidates in its pipeline.
The way AI is used:
For imaging analyses;
To select and design chemical compounds (ReChem);
To plan large, complex experiments (ReScreen);
To screen hundreds of thousands of drug compounds and cellular disease models (ReScreenRun);
To create digital mathematical signatures, or Phenoprints (ReRun);
To compute the effectiveness of each drug compound in their assays, as well as any unintended effects (ReAnalyze);
To model drug compound relationships (RePredict).
The Recursion Pharmaceuticals’ platform is a continuous, iterative loop of "biology and bits" which combines wet lab biology experiments that are executed automatically with machine learning algorithms computing the results in a cloud. The Recursion drug discovery platform is based on five million images of human cells every week requires built-for-purpose components
Analyse & Predict
To generate its datasets Recursion Pharmaceuticals is primarily focused on:
Data reliability and relatability:
By generating its own quality-controlled data, fit-for-the purpose of machine learning, Recursion is minimizing data noise to be able to ensure comparability of data.
Generalized assay framework for broad biology:
The technology is based on the principle of inducing the disease states and screening of various cell types together with healthy cells using fluorescent microscopy. By applying different substances, it’s possible to detect signals of potential drug-like molecules which return diseased cells to a healthy state, as well as potential side-effects.
Scale and scale again through automation and innovation:
Recursion’s software engineers, screening technicians and data scientists work closely to create a platform which allow to decrease the costs of drug discovery process. Recursion is trying to automatizate as much part of workflow as possible.
How OWKIN Uses AI in R&D?
Owkin is a predictive analytics company that was founded based on the belief that medical research must be collaborative, inclusive and protect privacy. Today, Owkin is building a global research network leveraging federated learning that brings data scientists, physicians, researchers and pharmaceutical companies together on a research platform that ensures data security and privacy.
Owkin is developing AI tools for medical researches to give patients access to safer and more effective therapies.
The way AI is used:
understand why drug efficacy varies from patient to patient;
enhance the drug development process;
identify the best drug for the right patient to improve treatment outcomes.
Owkin has developed a strong expertise in Hepatocellular Carcinoma (HCC), the most common form of primary liver cancer and the fourth leading cause of cancer death worldwide.
Owkin has created a unique research platform, and a portfolio of AI models and solutions.
The Owkin Loop is the heart of the Owkin Research Platform: it connects medical researchers with high-quality datasets from leading academic research centers around the globe. Owkin Loop is powered by the two main components of Owkin’s Software Stack: Owkin Studio, their machine learning platform, and Owkin Connect, their federated learning framework.
Owkin AI models
Owkin created a catalog of 30 live diseases models and has 40 additional models in the pipeline.These models differ from traditional black box models because they are built using interpreted AI, which allows the company to move further in research and identify biomarkers responsible for predictions. The discovery of new multimodal biomarkers is essential to identify new biological targets, optimize the design of clinical trials using patients subgroups, and identify patients eligible for a particular treatments.
DATA ENRICHMENT MODELS
Designed for translational researchers and pathologists, these models link histology to molecular markers.
OUTCOME PREDICTION MODELS
Designed for translational researchers, development executives, and commercial staff, these models can be implemented throughout the clinical drug development process to improve clinical trial design and evaluation, and to optimize product strategy.
PATIENT IDENTIFICATION MODELS
Designed for commercial and precision medicine leaders, these models help identify patients who will benefit from a particular treatment.
How LabGenius Uses AI in R&D?
LabGenius is the first biopharmaceutical company developing next generation protein therapeutics using a machine learning-driven evolution engine (EVA™).Their protein engineering platform integrates several cutting-edge technologies from the fields of machine learning, synthetic biology, and robotics.
The way AI is used:
To explore protein fitness landscapes;
To improve multiple drug properties simultaneously;
To produce DNA libraries.
LabGenius’ EVA Technology Platform
The EVA platform is able to design and optimize DNA libraries using AI, and consistently execute its production. These libraries encodes a host of evolved therapeutic proteins, whether these are antibodies or enzymes. For example, EVA can predict proteins resistant to proteases in the gastrointestinal tract. Eva consists a bank of liquid handling robots conducting experiments, and it is learning during the process, improving itself with each iteration.