Software detects hidden errors in spatial transcriptomics

Spatial transcriptomics is one of the most important methods in modern biomedical research. It enables the activity of cells within a tissue to be analyzed using RNA transcripts and assigned to individual cell types.

A research team at the Berlin Institute of Health at Charité (BIH) has now developed the software tool ovrlpy. It identifies previously difficult-to-detect sources of error in tissue analyses, such as overlapping cells or folds in tissue sections. The results were published in the journal Nature Biotechnology.

In many analyses, tissue samples are evaluated only in two dimensions. In fact, even very thin tissue sections with a thickness of five to ten micrometers – about one-tenth of a human hair – have a complex three-dimensional structure.

If these structures are interpreted only as flat planes, analysis problems can arise, for example due to:

• Overlapping of multiple cells
• Folds in the tissue
• Incorrectly assigned RNA signals

Such artifacts can result in transcripts not being correctly assigned to individual cells. This complicates the interpretation of the data and can influence further analyses.

The ovrlpy software tool analyzes the spatial distribution of RNA transcripts in three dimensions, detecting signal inconsistencies that indicate cell overlaps or tissue folds.

This approach allows potential sources of error in the vertical dimension to be identified that previously often went undetected. According to the researchers, analyses of various tissues and organs show that such overlaps occur more frequently than previously assumed.

"Ovrlpy helps us to identify these sources of error before they lead to false conclusions," says Dr. Naveed Ishaque, group leader for computational oncology in Roland Eils' Digital Health Department at BIH and last author of the study. He adds: "This creates the foundation for more robust insights in a wide range of disciplines, whether in cancer research, neurology or the development of personalised therapies."

With the increasing use of spatial technologies in biomedical research, the importance of reliable data is also growing. These methods provide detailed insights into the organization of cells within tissues. However, reliable results require a precise analysis of the complex tissue architecture.

The new ovrlpy tool can help to detect artifacts at an early stage and improve the quality of subsequent bioinformatic evaluations. It thus supports researchers in interpreting spatial data and can contribute to the development of new diagnostic approaches and therapeutic strategies in the long term.

MEDICA-tradefair.com; Source: Berlin Institute of Health in der Charité (BIH)