Home Publications Research Grants Inventions & Patents Awards Additional Research Activities Faculties & Institutions Research Areas

Automated multi-dimensional, high-content image-analysis for fluorescence-microscopy (A04) (SFB 796)

Third Party Funds Group - Sub project

Acronym: SFB 796

Start date : 01.01.2009

End date : 31.12.2012

Extension date: 31.12.2017

---

• Overview

   

• Publications

  (4)

• External partners

  (1)

Overall project details

Overall project

Overall project speaker:

Project details

Short description

To understand the molecular and structural basis of host-pathogen interactions, microbial effectors as well as their host targets must be studied in detail. As one key element for this research, fluorescence microscopy is an essential tool to evaluate, analyze and objectify the related experiments within microbiology, immunology, virology or similar disciplines. Specifically, image based experiments most often require the acquisition and evaluation of large amount of datasets, each consisting of several sub-images of various stains. Nevertheless, manual examination and analysis of such complex data is known to be prone to inter- and intra-observer errors. Also, due to limitations of resources, usually only a minimum subset of the depicted cells is analyzed manually, even though much more image data is actually available for most experiments and for more objective analyses. In order to improve reproducibility, objectivity and quantity of image based experiments, automated and semi-automated image-analysis methods must be utilized. Specifically, quantification of the cytoplasmic localization (C3), analysis of macrophage spreading and phagosome maturation (B6), cellular distribution and co-localization of different G protein-coupled receptors in HCMV infected cells (A6), analysis of the role of viral tax and p8 proteins on transmission of HTLV-1 (C6), interaction of herpes viruses and dendritic cells (B2) and nuclei dot counting (B1, B3) and intensity measurement (B5) require the determination of cell, nucleus or nuclear dot positions, related intensities and their spatial extension from different image channels, as well as 3D-texture analysis (C7). Additionally, for high-level analysis, a combination of the analyzed objects and information from various image channels is needed. In order to support these projects with automated, objective and reproducible image analysis, a novel image processing approach has been developed in the first phase of the CRC, allowing to automatically adapt the essential cell segmentation process to the fluorescent micrographs of a certain application. This approach has successfully been applied to various sub-projects (C3, B6, B1). With these methods and tools available and applicable, more complex and challenging image analysis problems arising in the cooperating projects can be addressed and supported in the upcoming period. Up to now, a diverse set of image processing routines for cell detection and separation based on single and dual channel image analysis routines were implemented, evaluated and applied to solve the related image analysis tasks. Nevertheless, to support the yet unsolved challenges in the cooperating sub-projects, which on the image processing level are related to the automated detection and splitting of touching, overlapping and even overlaying cells in a robust manner, phase contrast images shall be integrated into the analysis process to improve the splitting of overlapping cells. Similar wise, additional spatial and temporal information of cells must be considered as necessary extensions and included in the automated learning and image-analysis tools in the next step. Hence, the challenges of the upcoming funding period will focus on the research towards these multi-dimensional extensions for high-content image analysis.

Scientific Abstract

To understand the molecular and structural basis of host-pathogen interactions, microbial effectors as well as their host targets must be studied in detail. As one key element for this research, fluorescence microscopy is an essential tool to evaluate, analyze and objectify the related experiments within microbiology, immunology, virology or similar disciplines. Specifically, image based experiments most often require the acquisition and evaluation of large amount of datasets, each consisting of several sub-images of various stains. Nevertheless, manual examination and analysis of such complex data is known to be prone to inter- and intra-observer errors. Also, due to limitations of resources, usually only a minimum subset of the depicted cells is analyzed manually, even though much more image data is actually available for most experiments and for more objective analyses. In order to improve reproducibility, objectivity and quantity of image based experiments, automated and semi-automated image-analysis methods must be utilized. Specifically, quantification of the cytoplasmic localization (C3), analysis of macrophage spreading and phagosome maturation (B6), cellular distribution and co-localization of different G protein-coupled receptors in HCMV infected cells (A6), analysis of the role of viral tax and p8 proteins on transmission of HTLV-1 (C6), interaction of herpes viruses and dendritic cells (B2) and nuclei dot counting (B1, B3) and intensity measurement (B5) require the determination of cell, nucleus or nuclear dot positions, related intensities and their spatial extension from different image channels, as well as 3D-texture analysis (C7). Additionally, for high-level analysis, a combination of the analyzed objects and information from various image channels is needed. In order to support these projects with automated, objective and reproducible image analysis, a novel image processing approach has been developed in the first phase of the CRC, allowing to automatically adapt the essential cell segmentation process to the fluorescent micrographs of a certain application. This approach has successfully been applied to various sub-projects (C3, B6, B1). With these methods and tools available and applicable, more complex and challenging image analysis problems arising in the cooperating projects can be addressed and supported in the upcoming period. Up to now, a diverse set of image processing routines for cell detection and separation based on single and dual channel image analysis routines were implemented, evaluated and applied to solve the related image analysis tasks. Nevertheless, to support the yet unsolved challenges in the cooperating sub-projects, which on the image processing level are related to the automated detection and splitting of touching, overlapping and even overlaying cells in a robust manner, phase contrast images shall be integrated into the analysis process to improve the splitting of overlapping cells. Similar wise, additional spatial and temporal information of cells must be considered as necessary extensions and included in the automated learning and image-analysis tools in the next step. Hence, the challenges of the upcoming funding period will focus on the research towards these multi-dimensional extensions for high-content image analysis.

Involved:

Contributing FAU Organisations:

Funding Source