Artificial intelligence in the laboratory: current developments and applications

Artificial intelligence (AI) has become an integral part of modern laboratories. AI systems based on machine learning and algorithmic analysis are used in various scientific and medical disciplines to automate processes, increase efficiency, and gain new scientific insights.

In this article, we take a closer look at a few applications and venture a glimpse into the future.

AI is one thing above all: fast. Large amounts of data can be processed quickly and precisely, which significantly shortens diagnosis times and standardizes results, improving the reliability of diagnostics. For example, algorithms based on machine learning analyze high-resolution images of blood samples to detect differences between leukocyte types. These subtle differences are often crucial for the early detection of diseases. (Read more here)

At the same time, AI is also revolutionizing the prediction of laboratory results in clinical chemistry. By analyzing historical data, AI systems identify patterns that make it possible to predict future test results. This capability not only improves diagnostic accuracy, but also helps to avoid unnecessary tests and increases the efficiency of the entire diagnostic process. Especially in emergency situations, the fast and precise analysis capability of AI-based systems can be lifesaving. (Read more here)

In the pharmaceutical industry, the use of AI speeds up the process of screening and optimizing chemical compounds. Machine learning enables researchers to efficiently identify and optimize potentially effective substances from extensive chemical databases. This significantly shortens development times for new drugs and increases the likelihood that these compounds will be successful. The more accurate prediction of their efficacy and safety through AI thus has a profound impact on drug development. (Read more here)

The Fraunhofer Institute for Laser Technology ILT has developed the innovative LIFTOSCOPE system. The MIR LIFT process developed and patented by Fraunhofer ILT was integrated directly into the microscope system. A connected camera system generates one hundred high-resolution images per second, in which the AI recognizes the target cells using semantic segmentation.

The AI is trained to identify different cell types such as pluripotent stem cells, high-producer cells or immune cells. In the MIR LIFT process, the cells are then transferred individually to microtiter plates at a speed of up to 100 Hertz. "Depending on the cell type, up to 100 percent of the cells survive this process," explains Dr. Nadine Nottrodt, Head of the Biofabrication Group.

This process paves the way for the fully automated and highly efficient isolation of living cells. Dr. Nottrodt emphasizes that the results of the project so far indicate that it is possible to synchronize the cell LIFT process with the frame rate of the high-speed camera, which could enable a sorting speed of up to 100 cells per second. (Read more here)

Prof. Olga Golubnitschaja and her team from the University Hospital Bonn use AI for the early detection of stroke risks by analyzing tear fluid. As president of the European Association for Predictive, Preventive & Personalized Medicine EPMA (www.epmanet.eu) and head of the 3P Medicine research group at the University Hospital Bonn, she has been working on the paradigm shift from reactive to predictive, preventive and personalized medicine for several years. The aim is to achieve reliable interpretability of measurement results, non-invasive and painless sampling and cost-effective holistic applications that enable individualized prediction and targeted prevention even before the clinical manifestation of the disease in question.

AI is specifically used to make predictions about stroke risk from the metabolic patterns of tear fluid. These patterns are captured using multi-omics technologies that allow disease-specific signatures to be analyzed qualitatively and quantitatively. Prof. Golubnitschaja says of the innovative approach: "This provides us with the prerequisite for effectively counteracting multifactorial diseases such as ischemic stroke through predictive diagnostics followed by individualized treatment." The challenge lies in developing precise algorithms that can effectively process the huge amounts of data and provide reliable predictions.

The ongoing integration of AI in laboratories presents both opportunities and challenges. Trends include increasing automation of routine tasks, improving the predictive accuracy of diagnoses and personalized medicine based on individual patient data. However, challenges remain in the areas of data protection, the interpretability of AI decisions and the need for rigorous validation of algorithms to avoid misdiagnosis.

Progress in AI technology and its application in laboratory research and diagnostics continues to promise significant improvements. It is expected that AI will not only revolutionize research and clinical diagnoses but can also contribute significantly to the development of new therapeutic methods.