Pharma & Biotech
Case Studies
Newsfor Pharma & Biotech
Abbvie achieves 90% cumulative medication adherence among patients with schizophrenia using image recognition
Abbvie used AiCure’s artificial intelligence platform to visually confirm medication ingestion. This has facilitated continuous monitoring of patient treatment leading to better compliance and streamlining of clinical trials by reducing sample size.
Linked use case: Monitor patient prescription compliance
AstraZeneca improves internal management of its global data sources through organisation, search and information extraction using an AI platform
AstraZeneca implemented an AI-powered platform to extract information and automatically classify both its internal and external knowledge resources in order to aid information search and research and development efficiency within the company.
Linked use case: Scale and support data management and monitoring
AstraZeneca plans to crack down on online sale of counterfeit drugs in China using machine learning and natural language processing
AstraZeneca teams up with Tencent and Alibaba to strengthen the fight against counterfeit efforts. Alibaba will put traceability codes on drug packages and through their app patients will receive personalised recommendations and healthcare services. Tencent will work with AstraZeneca to identify online counterfeit drug sale using natural language processing to identify suspect language.
Linked use case: Identify and manage potentially fraudulent activity
BERG is attempting to identify genetic predisposition to certain conditions using machine learning with promising Phase I study results
BERG has developed a platform to swiftly analyze patient biology and identify biomarkers using machine learning. Their platform was used in Phase I study to discover molecular markers identifying patients more likely to benefit from the medicine, thus applying a precision medicine approach.
Linked use case: Identify and validate a molecule to target with a drug compound
BERG is developing targeted cancer drugs using machine learning with promising Phase I study results
BERG has developed a platform to swiftly analyze patient biology and identify biomarkers using machine learning. Their platform was used in Phase I study to identify patients more likely to benefit from the medicine, applying a precision medicine approach. They have used the platform to develop a new cancer drug and reduce chemotherapy-induced alopecia.
Linked use case: Identify and validate a molecule to target with a drug compound
Bayer aims to spot drug-associated side effects earlier with the use of machine learning, RPA and natural language processing
Bayer has partnered with Genpact to leverage artificial intelligence for its effort to process adverse events and spot drug-associated side effects earlier. The vendor is providing a solution that is able to optimise Bayer''s pharmacovigilinace practice for consumer health and pharmaceuticals. The Pharmacovigilance Artificial Intelligence (PVAI) solution works by automatically drawing AE data from unstructured and semi-structured source documents with the use of optical character recognition, robotic process automation, natural language processing, and machine learning.
Linked use case: Reduce side effects by collating patient data and optimising processes
Cambridge researchers develop a system to analyse cancer research papers to discover previously unexplored molecular biology links
Researchers at the University of Cambridge introduce a literature-based discovery (LBD) system to identify intermittently linked associations for cancer research in published literature. The LION LBD system uses convolutional neural networks and natural language processing to go through annotated databases of published research and come up with potential relations based on users searches. Results indicate at least a third of the proposed relations are useful and viable.
Linked use case: Re-examine data from historic research to discover new applications
Dana Farber Cancer Institute accelerates clinical trial recruitment by using machine learning for genome mapping to identify best candidates
Dana Farber Cancer Institute accelerates clinical trials by identifying genetic markers that respond better to particular drugs using machine learning.
Linked use case: Prioritise research and development projects
DayTwo improves blood sugar management through personalised diet recommendations based on gut microbiome analysis using machine learning
DayTwo is an app which provides users with their optimal diet based on their individual gut microbiome analysis. Ostensibly this is used to manage blood sugar levels and can help avoid developing prediabetes and type II diabetes, but can be used for a personalised approach to healthier eating in general. This is done through analysing biomaterial as well as eating and sleeping habits using a model developed with machine learning.
Linked use case: Create personalised food menu and diet
DeepMind develops a highly accurate machine learning method for predicting protein structures
DeepMind has developed a machine learning method to predict the 3D structure of proteins based on their genetic makeup. According to the company, the AlphaFold system is the most accurate methodology currently in the world for the prediction.
Linked use case: Enhance search process for new molecular structures
GlaxoSmithKline (GSK) plans to accelerate drug discovery as well as new applications for existing drugs using machine learning
GSK in partnership with Exscientia is exploring ways to accelerate drug discovery by identifying selective small molecules for 10 disease-related targets across multiple therapeutic areas. They plan to reduce the number of compounds required for synthesis in response to early stage research. They will then use machine learning to design molecules that fulfill lead and candidate requirements.
Linked use case: Identify existing drugs for improvement
GlaxoSmithKline (GSK) plans to reduce drug discovery to trial time from six years to 12 months using machine learning models to predict molecular behaviour
GSK in partnership with UC San Francisco attempts to discover new applications for previously developed compounds - both successful and unsuccessful, by analysing chemical and in-vitro biological data combined with publicly available data. This data is used to predict molecular behaviour and speed up clinical trials by targeting.
Linked use case: Re-examine data from historic research to discover new applications
Healx''s scientist predicts effectiveness of combinations of antibiotics using machine learning
Healx''s scientist Daniel Mason has developed a tool that uses machine learning to predict the effectiveness of combinations of antibiotics. Developed during his postdoctoral research associate position at the University of Cambridge under the supervision of Dr Andreas Bender, Chief Technology Officer at Healx, the tool aims to reduce the time and resources spent on experimental screening for new treatments. The Combination Synergy Estimation (CoSynE) is based on known structure of compounds and together with compound combination experimental screening data it can predict the activity of new compound combinations.
Linked use case: Predict potential adverse effects when drugs taken are combined
MIT scientists develop system that crowdsources data to speed up drug discovery using neural networks
Scientists at MIT have developed a system that lets pharmaceutical companies share their data to speed up drug discovery, while keeping them confidential. An artificial neural network (ANN), trained on 1.4 million drug-protein pairs that are known to both do and do not interact, identifies new drug-protein interactions. The system achieved that with 95% accuracy.
Linked use case: Optimise experimental efficiency through refining research process and operations
Novartis researchers train algorithm to identify different cell types to spot cancer in scans
Novartis has collaborated with PathAI to train a machine learning algorithm to categorise cell types and recognise cancerous ones. The system was trained on 400 pathology images from breast and lung cancer tissues from the Institute of Pathology at the University Hospital Basel. It had to identify cell types and if it spotted cancer, it had to predict a patient''s probability of surviving five years. The system was successful in categorising cells in five different types; lymphocyte, tumor cell, macrophage, plasma cell and fibroblast. The next step is to try to use the system for spotting information on scans and images that pathologists have missed or are not able to recognise.
Linked use case: Diagnose known diseases from scans, images, biopsies, audio and other data
Peptone accelerates protein research for drug discovery with a machine learning derived database
Peptone uses machine learning to predict structural stability for Intrinsically Disordered Proteins. This modelling has been compiled into a database of over 7,000 proteins, which can be integrated with other machine learning techniques to accelerate and enhance biochemistry research.
Linked use case: Leverage molecule database with metabolic stability data to elucidate new stable structures
Pfizer identifies new potential cancer treatments using IBM Watson
Pfizer has implemented IBM Watson to aid with cancer treatment research after it identified a treatment combination that the research team was also investigating separately. The platform is now being used for immune-oncology drug discovery research as a way to accelerate the process.
Linked use case: Identify existing drugs for improvement
Pharmaceutical company identifies warnings for non-Hodgkin’s lymphoma patients requiring change of treatment using machine learning
An undisclosed Pharmaceutical company has leveraged artificial intelligence to better understand non-Hodgkin’s lymphoma''s clinical progression and identify the best personalised treatment for each stage. The company applied machine learning to electronic health records (EHR) and other data to map out the warnings that indicate that patients need to switch to a later line of therapy. The team used an automated-feature-discovery (AFD) machine learning engine to test million hypothesis based on internal and external data. The system tried to find a statistically significant correlation between variables in patient data and transition to a later line of therapy. The technology enabled them to identify and isolate the variable combinations that predict transitions.
Linked use case: Predict personalised health outcomes to recommend individual treatment approach
Researchers at Macau University of Science and Technology develop a new model for disease classification in cases of limited labelled data with ~90% accuracy by combining logistic regression and semi-supervised learning
Due to lack of sufficient labeled data existing semi-supervised algorithms fall short in identifying related genes and classifying the disease. To overcome this researchers from Macau university have come up with a new logistic regression model combining active learning and semi-supervised learning which has achieved accuracy of above 90%.
Linked use case: Diagnose known diseases from scans, images, biopsies, audio and other data
Researchers at Stanford University develop an approach for modeling polypharmacy side effects with graph convolutional networks outperforming baselines by up to 69%
Researchers at the University of Stanford have developed a model to predict the side effects of using drug combinations, termed polypharmacy. Side effects are formulated as a multirelational link prediction problem in a two-layer multimodal network consisting of drugs and proteins. Decagon, the convolutional graph neural network model, is able to accurately predict the exact side effects of polypharmacy outperforming baselines by up to 69%.
Linked use case: Predict potential adverse effects when drugs taken are combined
Researchers at University of Lisbon accelerate drug discovery with the use of machine learning
Researchers at University of Lisbon, University of Cambridge, Friedrich-Schiller-University Jena, Federal University of Minas Gerais and Universidad de la Rioja are using machine learning to study how drug candidates and their targets interact to treat diseases. The algorithm that the team developed was able to identify that β-lapachone binds strongly to 5-lipoxygenase, an enzyme associated with human tumours.
Linked use case: Identify and validate a molecule to target with a drug compound
Researchers at the University of Glasgow develop platform to locate new molecules using machine learning
Chemists at the University of Glasgow claim that a new platform may help discover new molecules. The research team have trained an organic chemical synthesis robot to explore chemical reactions using image analysis machine learning software. In a demonstration, after exploring only 10% of the possible reactions the system predicted which combinations of starting chemicals should be explored to create new reactions and molecules with over 80% accuracy.
Linked use case: Enhance search process for new molecular structures
Roche plans to reduce time-to-market for oncology medicine by streamlining clinical trial process with machine learning
Roche aims to reduce the time taken to conduct clinical trial by assessing candidate eligibility from data and have acquired Flatiron Health who specialises in analysing EHR data, lab results, research, genomic data etc using machine learning to gather evidence and identify targeted trial candidates reducing time taken to launch medicine.
Linked use case: Optimise clinical trial design including patient selection
Sanofi Pasteur plans to make vaccines more effective by assessing biomarkers of influenza vaccination outcomes with machine learning
Sanofi Pasteur, the influenza vaccine global leader, is teaming up with BERG to gauge effectiveness of influenza vaccines among different groups of people. Sanofi will use BERG''s AI platform, to model and analyse vaccination outcomes to identify molecular signatures and potential biomarkers and their impact on Influenza vaccine immunological response. This will help them develop more effective vaccines for different groups.
Linked use case: Predict drug demand in different geographies for different products
Takeda and ConvergeHEALTH aim to better understand how treatment-resistant depression responds to medication using deep learning models
Data scientists from Takeda and institute ConvergeHEALTH by Deloitte are using insurance claims information such as diagnoses and prescriptions to analyse patient information. By looking at disease datasets like treatment-resistant depression they aim to identify factors that highly impact patient''s outcomes predictions. The analysis happens through deep learning models that go through patient histories to predict resistance to medication and benefits from switching.
Linked use case: Predict personalised health outcomes to recommend individual treatment approach