Artificial intelligence (AI) has been applied to numerous aspects of the life sciences, from disease diagnosis to drug discovery; in the first of this two-part blog series, we outline the impact of AI in Life Science and illustrate the various success stories of AI in Life Science.
Artificial intelligence, or AI, is the simulation of human intelligence processes by machines, and it is omnipresent; we encounter it daily, without necessarily being aware of it. AI encompasses the domains of Machine Learning (ML) and Deep Learning (DL) and is utilized in places such as smart homes, and driverless cars, and in cutting-edge research such as protein folding simulation.
We can find examples of AI within many sectors, including, but by no means limited to, business, education, finance, manufacturing, and, of particular focus to us at SciBite, the Life Sciences.
Life Sciences is the branch of science that is concerned with the study of living organisms and is responsible for the production of vast amounts of data. The innovation process, in particular, produces unbelievable amounts of data every day, enormous in all aspects of the 3V’s (volume, velocity, and variety), making the extraction of insights from such data cumbersome, so much so, that we need to turn to computers to help – enter AI.
But what specific tasks within Life Sciences has AI been applied to, and importantly what are some of the success stories? Are there any limitations that should be considered when utilizing AI in such a domain, and what do we see the future hold? The aim of this post is to give some thoughts and commentary on exactly these areas.
As the application of AI within the Life Sciences expands, so does our understanding of how and where it is best applied. Specific areas of application range from disease diagnosis all the way through to post-marketing surveillance, touching on every aspect of the drug discovery process along the way. Below, we present a purposely simplified grouping of these applications and provide some more demonstrative examples.
AI is used for predictive odelling utilizing clinical and demographic data to identify individuals who may be at risk of certain ailments. During the process of diagnosis image analysis is proving to be an attractive supplement used to analyze medical images coming from MRI, CT scans, or X-rays, particularly applied to neurodegenerative disorders, heart disease, and cancers, for example, spotting changes in breast mammograms. Work has also looked at how AI techniques may support less invasive monitoring approaches through virtual biopsies or via the monitoring of vital signs.
Unsurprisingly, AI has been applied throughout the drug discovery process, including in:
Manufacturers can leverage AI in many ways, including optimization of the process, monitoring QC, and ensuring the quality of products created match certain criteria, i.e., shape and size. Manufacturers can use AI to forecast demands; allowing for production to be scaled accordingly and can even look at limiting counterfeit drugs from reaching the market via smart packaging.
Much ongoing research focuses on how we can leverage AI to support personalized medicine, that is the development of bespoke treatment regimens; moving away from the outdated paradigm of ‘one size fits all‘. To support these efforts, AI models have been applied to large genomic datasets in the hope of identifying genetic markers associated with specific diseases.
Furthermore, the customer’s unique health history (dependent on the ability to have structured, digitized, data health records – see 7) have been utilized to make more informed treatment plans.
AI can be leveraged in many ways to support the design, recruitment into, and analysis of, clinical trials. Study design via identification of predictors of treatment response may can be developed using historical trial data. AI may also be used to expedite the identification of patients for enrolment based on their eligibility, in the identification of target populations and in estimating the ideal trial size. Once ongoing, AI may be used to support the automation of tedious tasks, such as data entry, improving efficiency, and reducing cost.
Monitoring the behavior of drugs after they have been rolled out is a crucial task for any pharmaceutical company, and as such many AI techniques have been deployed to support this. This work typically involves analyzing large amounts of textual data looking for potential safety and efficacy signals. Work tends to focus on a variety of sources, including toxicology reports, social media, or other real world data sources (RWD) and electronic health records (EHR).
Better management of EHR data can result in the patient and clinician being happier and money saved. RWD and social media has also been used for epidemic prediction.
Clearly, there are some very exciting examples of how AI is being used to enable us all to live better lives. Below are a few more specific success stories that we think are worth a specific mention.
Exscientia – in collaboration with Sumitomo Dainippon Pharma, reported the first AI-designed drug candidate to enter clinical trials in early 2020; DSP-1181 is being investigated as a treatment for obsessive-compulsive disorder. This new paradigm to drug discovery paved the way for others, with several companies having announced phase I trials with AI-designed drugs since, including Evotec, Insilico Medicine, and Schrödinger.
PoolBeg Pharma – made use of AI to identify more than one novel treatments for RSV infection; the approach utilised a vast amount of early-stage clinical data to prioritise repositioning candidates.
AlphaFold – developed by DeepMind, AlphaFold can accurately predict the three-dimensional structure of proteins from their AA sequence. A long-time elusive task, the model has been used to create open-source proteomic databases.1
Cleveland Clinic – in collaboration with IBM, is utilising AI to develop personalised healthcare plans for individuals. This effort makes use of mammoth amounts of health record data.
In part 2 of this blog, “Revolutionizing Life Sciences: The incredible impact of AI in Life Science”, we look at the challenges and future impact of AI in Life Science.
Joe Mullen, Director of Science & Professional Services. Holds a Ph.D. from Newcastle University in the development of computational approaches to drug repositioning, with a focus on semantic data integration and data mining. He has been with SciBite since 2017, initially as part of the Data Science team.
1. [Upcoming webinar] How important is subject matter expertise in Life Sciences when using technology and artificial intelligence? 29 March 15:00 BST.
2. [Blog] Why Use Your Ontology Management Platform as a Central Ontology Server, read more.
3. [Blog] SKOS in CENtree: Further support in our latest 2.1 release, read more.
4. [Blog] Talking to TERMite – introducing the SciBite scripting suite, read article.
As discussed in part 1, Artificial intelligence (AI) has revolutionized several areas in life sciences, including disease diagnosis and drug discovery. In this second blog, we introduce some specific text-based models whilst also discussing the challenges and future impact of AI in Life Science.Read
Raw data has the inherent characteristic of being unstructured with potential quality issues such as inaccurate, incomplete, inconsistent, and duplicated. Therefore, it must be processed before it can be used for subsequent analysis and confident data-driven decisions. This is where ontologies come into play.Read
Get in touch with us to find out how we can transform your data
© Copyright © 2023 Elsevier Ltd., its licensors, and contributors. All rights are reserved, including those for text and data mining, AI training, and similar technologies.