Introduction

Biobanks are essential for personalised medicine. They provide diverse biological samples and data necessary for tailored treatments. By studying this data, researchers can develop personalised disease prevention, diagnosis, and treatment approaches, ensuring better patient outcomes.¹

Biobanks produce big data by collecting vast amounts of diverse samples and related information. This data is complex and continuously growing, requiring innovative methods for storage, analysis, and usage. Big data from biobanks helps discover new preventive methods and optimise treatments.¹

Artificial intelligence (AI) analyses big data efficiently, handling its volume, variety, and velocity. It helps predict health outcomes, identify risk factors, and enhance patient care. By processing large datasets from biobanks and other sources, AI can support personalised medicine and better healthcare decisions​.¹

Harnessing AI for efficient biobank data analysis

Biobanks collect a massive amount of data, but we lack reliable ways to analyse it. The data is too vast for traditional visual analysis or simple statistics. AI and its subfield, machine learning, can solve this by efficiently handling and finding patterns in large datasets. Unlike traditional methods, AI and machine learning do not need explicit programming to perform specific tasks. They can independently detect and analyse data patterns.²

In biobanks, AI helps in many ways. It understands consent forms and answers questions online. It also predicts how patients might do and helps with real-time experiments.³ 

Thus, using data from biobanks, AI and machine learning can predict the properties of drugs and potential treatments. Overall, AI and machine learning offer a promising solution to managing the growing volume of complex data.²

Advancing precision medicine through AI in genomic analysis

One way precision medicine uses AI is by applying it to genomics. AI helps combine algorithms to analyse and predict diseases more efficiently. It focuses on gene activity, DNA changes, and other genetic alterations. Machine learning and deep learning (another part of AI) are used to create models that speed up data analysis using genomic and other data from biobanks. Integrating different types of data is crucial for capturing the full complexity of each approach.² 

While progress has been made, more research is needed to find the best AI methods for this work. Integrating multiple types of data can help us understand diseases better than using genomics alone.²

United Kingdom biobank

UK Biobank is a massive health database with detailed, anonymous genetic and health information from over 500,000 UK participants. Regularly updated, this data is accessible to approved researchers worldwide studying common and serious diseases. UK Biobank has helped make several key scientific discoveries that advance modern medicine.^4,5

Since 2006, UK Biobank has gathered vast amounts of biological and medical data from people aged 37-73 across the UK. Participants regularly provide blood, urine, and saliva samples, along with lifestyle details; all linked to their health records. It collects extensive details about participants, including questionnaire responses, physical measures, genetic information, and health follow-ups. This helps researchers understand how diseases affect people.

Researchers from academic, charity, government, and commercial organisations use this data for public health research. UK Biobank is the most detailed long-term health study in the world, helping scientists understand various diseases. UK Biobank continues this work by using new technologies and regular health follow-ups.

Developing this resource required input from scientists, managers, and legal and ethical experts. It is available for any genuine researcher to use for public health research. Moreover, the wide consultation, Ethics and Governance Framework, and oversight ensure UK Biobank meets ethical and legal standards. This rigorous approach reassures regulatory bodies and protects participant interests.

AI-based risk scores for predictive, personalised medicine using UK Biobank data

AI promises big advances in health science. UK Biobank’s vast data helps researchers improve disease prevention, diagnosis, and treatment. AI can process layers of health information, including genetics and lifestyle, to predict disease risk and personalise prevention and treatment.

In 2023, UK Biobank proposed using its detailed data to develop AI tools for predicting diseases like diabetes, heart disease, and COVID-19. These tools could reduce NHS costs by improving disease prevention and management. The project, planned for at least three years, aims to validate these tools across different populations and test them in clinical trials.

Applications of AI in disease detection with biobanking datasets

In 2022, a scientific review analysed multiple studies which focused on AI’s role in biobanks. The review identified various applications of AI in disease detection with biobanking datasets.³

• Alzheimer’s disease detection
  • One study analysed 500,000 patients from the UK Biobank and revealed an 82.4% accuracy. Biobank data also helps AI predict age-related macular degeneration (AMD) 
  • Deep learning uncovers brain patterns, examines social brain areas, and even predicts social traits like isolation. AI uses MRI scans contained in biobanks to spot Alzheimer’s disease
• Cardiovascular diseases
  • AI forecasts heart disease risks using massive biobank datasets. One study used an auto-prognosis tool to fine-tune machine learning model features for sharper predictions
• Chronic diseases
  • AI predicts diabetes, obesity, and cancer risks by analysing biomedical samples. A study of 1,000 patients from Qatar biobanks used machine learning to assess chronic disease risks 
  • In another study, AI pinpointed risk factors for diabetes and obesity 
  • A machine learning algorithm predicted breast cancer treatment side effects with 75.93% accuracy by analysing 695,227 genetic variations from UK biobanks 
  • AI also predicts hypertension and keeps tabs on behaviours like walking and sleeping to keep adults healthy
• Disease subtype classification
  • AI improves diagnosis and treatment by classifying disease subtypes and related biomarkers. Both humans and machines struggle with noisy biomedical data. A reviewed study introduced a new machine learning approach using UK and Atlas biobank data to classify disease subtypes
• Pandemics
  • AI accurately assesses COVID-19 risks and tracks its progression. It predicts the risk of death, hospitalisation, and ICU admission based on COVID-19 data 
  • A study used machine learning models to estimate mortality risk in COVID-19 cases from the UK Biobank, factoring in issues like kidney failure and pneumonia

Advancing treatment for mental disorders: biobanks, AI, and personalised medicine

Mental disorders are a big deal globally - they're major causes of long-term illness, disability, and sadly, even death. Precision medicine in psychopharmacology aims to tailor treatments based on each person's genetic makeup, environment, and lifestyle. It tries to use real-world data from biobanks to create prediction models for how treatments will work for mental disorders. This way, researchers are not just relying on data from clinical trials, but on a broader range of info.⁶

Ethical considerations

Ethics are crucial in AI because biases can sneak into models from how data is collected or used. These biases stem from who is in the data and how it is used in AI systems. Nowadays, there are no clear rules for reporting and comparing AI models. It is important to identify these biases to help guide researchers and doctors.⁷

As AI becomes more essential in healthcare, ensuring it makes fair decisions is vital. Responsible AI should be clear, understandable, and accountable. This is especially important because AI can improve patient care through precision medicine. Ignoring AI in healthcare could be seen as not scientific and wrong.⁷

Implementing AI in biobanks can find new biomarkers, develop better tests, and choose treatments that are better for the environment and cheaper. However, AI needs physicians to make sure the data and images are right for training its algorithms. The future is in AI and physicians working together to give the best healthcare possible.⁷

FAQs

What are biobanks?

Biobanks collect, store, and share biological samples like blood, urine, and tissues, along with associated data like clinical (health records) and imaging information (MRI / CT and X-ray scans etc). They support research and personalised medicine by enabling scientists to study diseases and develop targeted treatments.¹

Biobanks come in two main types:³ 

• Population-based: focused on broader population health, they study various diseases
• Disease-oriented biobanks: focused on specific diseases, they store medical information and genetic samples

What is personalised, precision medicine?

Personalised, precision medicine tailors medical treatment to the individual characteristics of each patient. It aims for the right treatment at the right time for the right person. It considers factors like genetics, environment, and lifestyle to prevent, diagnose, and treat diseases more effectively. Unlike regular medicine, precision medicine recognises that treatments do not work the same for everyone. Biobanks play a crucial role by providing the necessary data and samples.^1,2

What is big data?

Big data refers to extremely large datasets that are complex and grow rapidly. These datasets require advanced tools for storage, analysis, and interpretation.AI helps manage these processes.¹

What is artificial intelligence? 

AI is like a computer that learns and thinks like humans do. AI systems learn and adapt by interacting with training data, unlike traditional algorithms with fixed rules. Within AI, machine learning and deep learning techniques get smarter with experience. Essentially, AI involves machine learning and deep learning computer algorithms using data to make decisions or predictions.^1,3

Summary

Biobanks and artificial intelligence (AI) are changing up personalised medicine. Biobanks are collections of biological samples and data that scientists use to figure out personalised treatments. Personalised, precision medicine tailors treatments based on genetics, environment, and lifestyle and is closely tied to this. AI comes in by crunching all this data efficiently, predicting health outcomes and making treatments better tailored for each person. They team up to unlock new ways to tackle diseases, from diabetes to cancer, using big datasets. In addition, researchers try to make sure AI plays fair and stays accountable.