Passionate about AI in Health

Pediatricians and computer scientists join forces at the Helmholtz Munich Institute for Diabetes Research to identify children with type 1 diabetes before they develop symptoms. With the help of advanced statistics and machine learning, the researchers can predict the progression of early-stage type 1 diabetes and pinpoint the effective timeslot for intervention with potential preventive measures.

Our health and the environment around us are profoundly intertwined. But how do air pollution and other environmental factors interact with each other and determine our health for each individual and the neighborhood? Furthermore, how can artificial intelligence (AI) and machine learning (ML) assist us in comprehending and identifying the impending effects?

Researchers at Helmholtz Munich have made significant progress in advancing high-resolution optoacoustic imaging for clinical use. Their innovative deep-learning framework, known as DeepMB, holds great promise for patients dealing with a range of illnesses, including breast cancer, Duchenne muscular dystrophy, and inflammatory bowel disease. Their findings have been now published in Nature Machine Intelligence.

A collaborative endeavor between computer scientists and diabetes researchers at Helmholtz Munich has yielded novel insights into the mechanisms underlying type 1 and type 2 diabetes. This collaboration has resulted in the creation of the first mouse islet atlas (MIA). Leveraging the power of machine learning, the team of scientists integrated single-cell datasets to reveal the molecular alterations that occur during the progression of diabetes and to highlight the distinctions between type 1 and type 2 diabetes. Their findings have been published in Nature Metabolism.

Colorectal cancer (CRC) ranks second in leading causes of cancer-related deaths globally, according to the WHO. For the first time, researchers from Helmholtz Munich and the University of Technology Dresden (TU Dresden) show that artificial intelligence (AI)-based predictions can deliver comparable results to clinical tests on biopsies of patients with CRC. AI predictions can speed up the analysis of tissue samples, resulting in faster treatment decisions.

Researchers developed a new method called wildDISCO that uses standard antibodies to map the entire body of an animal using fluorescent markers. This revolutionary technique provides detailed 3D maps of structures, shedding new light on complex biological systems and diseases. WildDISCO has the potential to transform our understanding of intricate processes in health and disease and paves the way for exciting advancements in medical research.

Computer scientists at Helmholtz Munich developed a generative model named scPoli (single-cell population level integration), that enables multi-scale representations of cells and samples. The model performs data integration of high-quality large-scale datasets of single cells to create valuable single-cell reference maps of the human body, so-called single-cell atlases, for medical research.

Can a human organ be mapped on a single-cell level to learn about the functionality of each individual cell? And can we learn how different these cells are from person to person? Researchers have taken up this challenge and developed the Human Lung Cell Atlas using artificial intelligence (AI)-based techniques.

Unlike other forms of blood cancer, acute myeloid leukemia (AML) cannot currently be treated with CAR-T cell immunotherapy. The reason is that specific molecular targets with which certain immune cells could specifically target AML cells are lacking. Researchers have now succeeded in discovering such targets.

Researchers created a novel deep learning method that makes automated screenings for eye diseases such as diabetic retinopathy more efficient. Reducing the amount of expensive annotated image data that is required for the training of the algorithm, the method is attractive for clinics.

Researchers developed a novel artificial intelligence algorithm for clinical applications called “scArches”. It efficiently compares patients’ cells with a reference atlas of cells of healthy individuals. This enables physicians to pinpoint cells in disease and prioritize them for personalized treatment in each patient.

Researchers from Helmholtz Munich have reached a significant milestone in precision medicine with their new publication on the 'Oncology Biomarker Discovery (OncoBird)' framework. This innovative approach aims to systematically identify actionable biomarkers for cancer treatment in clinical trials. By analyzing the molecular and biomarker landscape of randomized controlled trials, OncoBird successfully pinpoints predictive biomarkers in metastatic colorectal carcinoma patients, distinguishing which treatment option - cetuximab or bevacizumab - would be more effective based on specific genetic alterations and tumor subtypes.

Human health is closely connected to the health of our environment. So far, the complex ways in which they influence each other are not fully understood. The concept of One Health recognizes this interconnectedness of human, animal, and environmental health. Researchers show, how using real-time genomic analysis can benefit the concept of One Health. A combination of statistical and AI methods allows for analysis of genomes in real-time. This helps to understand health of ecosystems in a more detailed and timely manner.

CAMPA, a new deep-learning framework, that allows for analysis of subcellular organization in healthy and perturbed cells using high-resolution fluorescence microscopy data has now been introduced in Nature Methods. This tool can facilitate the analysis of high-throughput screens and accelerate researchers' ability to gain insight from complex datasets.

Predicting the effects of drugs regarding dosage, timing, possible drug combinations and other types of intervention such as gene knockouts can be experimentally difficult and very time-consuming. Nonetheless, this is one of the most important tasks in drug development and pharmaceutical research. A team of researchers developed the first open-source computational model based on generative AI to predict, interpret, and prioritize perturbations in cells. This can speed up the testing process and can serve as a guide for experimental validation.