Analytics for Spatial Biology: DNA and RNA Imaging

Gradient boosting machine (GBM) learning is applied to analyzing DNA and RNA images created using AI and multiple analytical techniques. AI and GBM hold promise for simplifying and standardizing discovery for spatial biology. During a lecture at the 2024 Analytica conference in Munich, Germany, several experts spoke about this technology.

To begin this session, Denis Schapiro, from the University Hospital Heidelberg and Heidelberg University in Germany, presented "From oncology to cardiology: Spatial omics technologies for topographic biomarker discovery," emphasized the development of the histoCAT software toolbox designed for highly multiplexed image analysis, particularly from imaging mass cytometry (IMC). Imaging mass cytometry (IMC) is a technique that displays the spatial distribution of proteins or other biomolecules within tissue samples. These images are generated by combining mass spectrometry using metal-tagged antibodies, enzymatic methods, or by using fluorescence spectroscopy. Schapiro introduced histoCAT's advanced machine learning (ML) approaches and its integration with the modular computational pipeline, MCMICRO, enabling proteomic and transcriptomic analysis across various spatial technologies. Additionally, Schapiro discussed the highly multiplexed tissue imaging (MITI) standard and a spatial power analysis framework to enhance experimental design strategies, demonstrated through data processing related to myocardial infarction.

The second talk by Ralf Jungmann, of LMU Munich and Max Planck Institute of Biochemistry in Germany, gave a lecture titled "From DNA Nanotechnology to Biomedical insight: Towards Single-Molecule Spatial Omics," outlined advancements in DNA-PAINT software for converting standard fluorescence microscopy into a spatial omics tool. The analytical toolkit used for spatial omics typically includes techniques such as imaging mass cytometry (IMC) and spatially resolved RNA sequencing (spatial transcriptomics). These methods enable the simultaneous measurement of molecular and spatial information within tissue samples, facilitating the study of cellular heterogeneity, interactions, and spatial organization in biological systems. Jungmann introduced improvements achieving sub-nanometer spatial resolution and spectrally unlimited multiplexing, along with strategies to increase imaging speeds in DNA-PAINT. Furthermore, he presented cell surface receptor quantification techniques and their potential for therapeutic applications.

In the third session talk, Manuel Liebeke, from the University of Kiel, presented, "Deciphering Metabolism in Host–Microbe Interactions with Mass Spectrometry Imaging and Microscopy," discussing the use of mass spectrometry imaging (MALDI-MSI) and spatial metabolomics fluorescence in situ hybridization (metaFISH). These methods are used to study host–microbe interactions by allowing direct and simultaneous mapping of diverse metabolites within biological tissues. Liebeke showcased metaFISH's ability to assign spatial distribution of metabolites to specific microbiome members at single-cell resolution, providing insights into metabolic interactions in dynamic environments. Through metaFISH, precise localization of bacteria, host cells, and associated metabolites in animal tissues was demonstrated, enhancing understanding of metabolic interactions.

The final presentation of this session was given by Martin Seifert, of 10X Genomics in Leiden, the Netherlands. The talk was titled, "New Possibilities for the Discovery of Disease Relevant Information. Gaining a new Picture of Biology with Single Cell and Spatial Analyses." This talk highlighted the integration of single-cell sequencing, spatial transcriptomics, and targeted in-situ analyses for disease tissues. Seifert emphasized their potential in elucidating molecular patterns crucial for understanding various disease processes, exemplified through data processing related to formalin-fixed paraffin-embedded (FFPE) cancer samples. The lecture highlighted the complementary nature between different analytical technologies, providing new insights into disease-relevant processes. Mass spectrometry (MS) is most often employed to analyze the proteome and metabolome of single cells.