Group Members

MSc Students

• Michał Traczyk - Non-Parametric Clustering in a Probabilistic Graphical Model Describing Spatial Cancer Immunology
  

  Michal developes a non-parametric clustering method within a probabilistic graphical model to analyze multimodal cancer data from the IMMUcan consortium. The big goal is to use the Chinese restaurant process for cell cluster identification and extends frameworks like MOFA+ to integrate high-resolution imaging and sequencing data.

• Gabriel Kiciński - Modelling and analysis of next-to-tumor tissue on spatially-resolved data to understand pre-tumor markers
  

  This research analyzes spatial data from cancer and adjacent tissues to identify early cancer signs and compare cancer-specific features in seemingly healthy areas. Using graph-driven methods and tools like CellCharter, squidpy, and SpaceLet, the study aims to develop new solutions for comparing cellular neighborhoods while preserving their spatial properties.

• Stanisław Janik - Detecting and classifying glomeruli structures in histopathological images
  

  This work focuses on applying and evaluating cell segmentation and glomerular sclerosis classification models on PAS-stained kidney images, with an emphasis on explainability and spatial features of glomeruli. The study aims to develop a customized classification model and analyze which features influence the model in collaboration with expert pathologists from Uniklinik Freiburg.

Former Supervisions

PhD Students

• Barbara Poszewiecka (co-supervised with prof. Anna Gambin) - Computational Methods for the Analysis of Chromosomal Rearrangements
  

  Barbara's dissertation focuses on algorithms for genome assembly and interpreting genome architecture changes caused by structural rearrangements. It presents innovative methods for analyzing third-generation sequencing data, speciation timing, and complex chromosomal rearrangements, as well as a web server for clinical interpretation of structural variants.

MSc Students

• Joanna Krawczyk - Statistical modeling and analysis of multiplex immunofluorescence data
  

  Modern cancer diagnosis benefits from multiplex immunofluorescence (mIF) technology, which provides high-resolution data describing the tumor immune microenvironment (TIME) for precise analysis. Joanna's work combines classical image processing techniques and graph-based methods to analyze spatial clusters of cells, aiming to detect clinically significant structures and identify patterns of infiltration that distinguish patient groups.

• Aleksandra Możwiłło - Graph-based spatial methods to infer immune infiltration patterns in cancer multiplex immunofluorescence imaging
  

  Aleksandra's work focuses on spatial analysis of immune cells within tumor tissue using multiplexed immunofluorescence data from the IMMUcan consortium. The main objectives are to introduce a novel clustering method for immune-isolated regions in the tumor microenvironment and develop a Python package for spatial tumor analysis.

• Jakub Białecki - IsletWeb - a webservice supporting graph-based modeling and analysis of multiplex immunofluorescence data
  

  With the success of immunotherapy, new methods and tools are being developed to better understand the tumor immune microenvironment (TIME), which is a key diagnostic and prognostic factor for cancer patients. Jakub aims to build a web service for processing and analyzing multiplex immunofluorescence (mIF) data, integrating the TumorIslet package for visualizing spatial relationships in tumor tissues and proposing new methods for characterizing spatial dependencies to group patients accordingly.

Former BSc Students

• Tomasz Karolewski - Analysis of the COVID-19 development dynamics in European countries
  

  Tomasz discusses statistical methods for processing and analyzing COVID-19 infection data from European countries. He introduces classic techniques like smoothing and clustering, as well as an original method to model infection waves and predict their end using the Random Forest algorithm.