News & Views
Laboratory AI - How Does It Help Scientific Research?
Dec 07 2021
Artificial intelligence (AI) has reimagined the limits of what’s possible in science laboratories. From cancer research and genetic biology to drug development and food quality control, AI technologies are used across a myriad of scientific research fields.
Want to know more about how laboratory AI has emerged as an invaluable tool for the scientific research realm? Below, we take a look at the exciting new technologies being used across a variety of research fields.
The advent of artificial intelligence
The origins of artificial intelligence can be traced to the 1940s, when English mathematician and computer scientist Alan Mathison Turing began to question if machines can think. He eventually developed an experiment called the Turing Test, designed to determine if machines are “intelligent” based on responses to questions.
If the machine could mimic human responses and convince a person it was a fellow human, it passed the test and was considered intelligent. Turing colloquially referred to the test as the “imitation game” and is widely considered the grandfather of AI.
Over the past few decades, artificial intelligence has become commonplace in laboratories around the world. Today, it’s used as an umbrella term to describe an arsenal of technologies used to replicate human intelligence and carry out complex tasks.
Modern advances in AI technologies
AI has come a long way since the Turing era. Advances in AI technologies have allowed analysts to collect a huge amount of data, with next-generation software and algorithmic techniques used to transform this data into exploitable information. Here are some of the biggest AI breakthroughs from the past few decades.
Convolutional neural networks
Convolutional neural networks, also known as CNNs or ConvNets, allow AI systems to analyse visual imagery and solve problems by assigning importance and making differentiations. The complex networks are made up of neurons and inspired by the connectivity patterns and neural pathways that exist in the human brain, specifically the visual cortex.
Reinforcement learning sees AI systems explore a variety of different scenarios and possibilities. Credit is then assigned to different options based on performance.
Transfer learning sees previously learned concepts used in new, unknown scenarios.
Generative adversarial networks
Developed by AI scientist Ian Goodfellow in 2014, generative adversarial networks (GANs) stake a pair of neural networks against each other. By creating a ‘zero-sum game’ scenario, researchers can force machines to learn new techniques and generate valuable data.
AI transforms scientific research sector
For the scientific research sector, artificial intelligence has been a key driver of many breakthroughs. Researchers across a broad range of disciplines rely on the tool to extract and analyse data, with information used to gain insight, detect patterns and predict outcomes.
Here’s a glance at how AI is currently being used in some of the world’s top scientific research laboratories:
AI in genomic research
Artificial intelligence has emerged as an invaluable tool for genomic research, with data used to predict complex protein structures. By predicting the different shapes protein structures acquire and how they affect the human body, researchers can develop a deeper understanding of disease development. AI-derived data is also used to improve diagnosis and develop new treatments. As well as posing technical challenges, using genetic sequences to predict the shape of a protein is incredibly labour intensive.
The 100,000 Genomes Project
The now- finalised 100,000 Genomes Project overseen by Genomics England demonstrates just how complex genetic sequencing can be, with the team citing data as a major challenge. Using data from National Health Service patients, the project sequenced whole genomes with the goal to record more than 100,000 known structures and develop a better understanding of common cancers, as well as rare and infectious diseases.
The first hurdle was variant calling, which involves comparing the millions of possible differences between a reference genome and the unique genome of the patient. Annotation, the process of deriving meaning and relevance from each difference, was also an enormous challenge. Rather than manually attempt these complex tasks, researchers used artificial intelligence and machine learning techniques to analyse DNA sequences and predict three-dimensional proteins structures.
“Automating this process - creating the Genomics England pipeline - so that it took weeks rather than years, was very difficult,” reads the Genomics England website.
Eagle Genomics and the Earlham Institute
A recent partnership between Eagle Genomics and the Earlham Institute further demonstrates the transformative potential of AI in the scientific research sector. The collaboration will focus on developing critical tools to access and analyse complex microbiome data.
“Eagle Genomics’ industry-leading open platform is helping drive the digital reinvention of life sciences R&D across numerous industries, from AgBio and food through to healthcare, personal care and beauty,” says Eagle Genomics CEO, Anthony Finbow. “Our partnership with the Earlham Institute is a clear demonstration of our commitment to enable a radical, global shift in microbiome science-led innovation. By harnessing nature - including its complex networks and associated multi-dimensional data - we can help tackle the world’s Grand Challenges while unlocking significant economic potential.”
Combating climate change
Climate change is a pressing issue and AI is being used to help environmental scientists unlock valuable data and rally for action. From monitoring air quality patterns to recording changes in average monthly rainfall, environmental science is heavily reliant on data. Complex systems are fundamental to analysing this data and using it to make informed decision, at local, national and global levels.
With so much data and so many different parameters, patterns and trends to factor in, manually generating meaningful data can be a challenge. Artificial intelligence bridges the gap between the slew of information available and the potential impacts environmental changes may have on the planet.
Predicting patterns and planning accordingly
For example, researchers at the University of Cambridge are currently investigating how climate change will affect cotton production in Egypt over the coming decades. Scientists are drawing on local and international statistics, with AI systems used to transform raw data into useful information.
In the United States, global private investment firm TPG channelled US$100 million into an AI-driven system designed to predict weather patterns with extraordinary accuracy. The data-rich system was developed by Climavision, a Kentucky-based forecast modelling start-up with big potential.
“As weather patterns become increasingly unpredictable and volatile due to climate change, the need for higher-quality regional and hyper local weather data has never been more pronounced,” says Chris Goode, Climavision CEO and Co-Founder. “Climavision’s increased coverage and improved weather information enables earlier and more accurate weather forecasts that can save lives, limit business disruption, and improve the lives of people and communities across the country.”
Powering renewable energy
As the planet pledges to combat climate change and make the shift to renewable energy, experts predict AI will play an important role. Technologies are already being used to manage decentralised renewable energy grids and distribute supply.
From an operational perspective, AI is used to monitor the condition and performance of renewable energy infrastructure, including wind turbines, solar panels and hydroelectric machinery. By improving efficiency, minimising downtime and reducing maintenance costs operators can keep prices as low as possible and fast-track the global rollout of green energy.
Astronomical research and space exploration
Artificial intelligence isn’t just fuelling scientific research here on Earth. Machine learning is also being used to detect patterns in astronomical data and develop a deeper understanding of outer space. For example, NASA uses AI to detect existing data sets to detect magnetised rotating compact stars known as pulsars. Machine learning is also used to map the physical properties of stars, profile supernovae and categorise galaxies.
During NASA’s famous Kepler Mission, a purpose-built AI system was used to remove distorted data generated by the space telescope’s thrusters. This allowed researchers to focus exclusively on astronomical data generated by stars and planets.
Beyond the laboratory, artificial research is being used to supercharge asset management strategies. A recent feasibility study pioneered by British-owned procurement solutions company The NBT Group, industrial software developer Senseye and Northumbria University secured a £250,000 to investigate the use of intelligence-based technologies to connect predictive maintenance software with advanced stock management systems. The goal is to minimise downtime, improve performance and harness the potential of emerging smart technologies.
“We’re always looking to innovate and partner with similarly minded organisations,” says Senseye Chief Technology Officer, Robert Russell. “This Innovate UK funding enabled us to deepen our relationship with both The NBT Group and Northumbria University in an exciting opportunity to bring together very different aspects of asset management that is only made possible by advancing digitisation of the manufacturing sector. The feasibility study is accelerating our vision for advanced technologies and generating new business opportunities.”
The future of AI and scientific research
Artificial intelligence is continually evolving, with new developments always pushing the boundaries of scientific research. Moving forward, it’s certain AI will continue to play a pivotal role in advancing modern science and revolutionising research across a huge range of disciplines, from climate change and sustainability to astronomy and medicine
In This Edition Chromatography Optimising Viral Vector Purification Strategies with Multimodal Chromatography Key UHPLC Characteristics Required for High throughput LC-MS New Low Volum...
View all digital editions
Dec 20 2022 Shanghai, China & Online 21 November 2022 to 13 January 2023
Jan 25 2023 Tokyo, Japan
Jan 30 2023 Dubai, UAE
Feb 01 2023 Tokyo, Japan
Feb 06 2023 Dubai, UAE