Cell Atlas with AI improves the diagnosis of inflammation

The inflammation It is behind many common diseases, from infections to immune disorders or some types of cancer. However, understanding how exactly it acts in the human body remains a challenge. Now, the National Center for Genomic Analysis (CNAG), together with the École Polytechnique Fédérale de Lausanne, has taken an important step to clarify this: the first cellular atlas of inflammationwhich uses artificial intelligence to improve the diagnosis of these pathologies.

It is a large-scale database that gathers information from more than 6.5 million blood cells coming from more than 1,000 peopleboth healthy and patients with 19 different diseases. The work has been published in the magazine Nature Medicine.

The objective has been to observe what happens when inflammation is deregulated, either in infections such as Covid-19immunological diseases such as psoriasis wave rheumatoid arthritisor in cancers such as breast or colorectal. Until now, many of these processes remained poorly understood.

The persistent inflammation caused by rheumatoid arthritis can eventually damage bones, ligaments, and tendons. Source: Europa Press.

One of the key contributions of the study is to treat the cells themselves as biomarkersthat is, as signs that allow the identification of a disease. As cells circulate through the body, they collect information about what is happening in it, and this “footprint” can be analyzed.

Researchers have studied different cell states and the genes related to inflammation that coordinate the immune response: those that activate defenses, guide cells to damaged areas or help fight harmful agents.

Thanks to this analysis, the atlas identifies molecular indicators that function as guiding signs. These allow us to more accurately distinguish some diseases from others and classify patients according to their inflammatory profile, something key to adjusting treatments. Each disease presents a particular combination of signals, which opens the door to finer and more personalized diagnoses.

Artificial intelligence learns from inflammation

The project incorporates a model of generative artificial intelligence able to learn from atlas data. This system can recognize patterns in genetic activity and cellular states, and apply them to future patients.

According to the researchers, this approach lays the foundation for a precision medicine tool which could speed up diagnoses and help decide on treatments that are more tailored to each case.

The atlas is now available as open source resource and has been tested with real samples as diagnostic support. The next step will be to define data quality and standardization protocols, an essential requirement so that this knowledge can be safely integrated into routine clinical practice.