AI-assisted analysis for cancer drug development (KASPAR)

The project goal is to enable the use of protein profiles from samples with few cells for drug development. For this purpose, a workflow consisting of automated sample preparation, AI-supported image analysis and data evaluation as well as mass spectrometry will be developed and validated.

Tumours are heterogeneous entities and may, for example, contain isolated tumour cells with stem cell-like properties, which are considered to be the main reason for resistance to a drug or the occurrence of a relapse. In order to be able to map this heterogeneity, single-cell analyses are increasingly becoming the focus of attention. In addition, xenograft models (PDX) derived from patient material are increasingly used in drug or clinical method development due to the large number of experiments required. Here, human tumour cells are cultivated in mice. PDX correspond as far as possible to real patient samples, but without limitations in terms of sample material.

In order to characterise single cells, mainly the DNA contained in them is currently analysed. However, the DNA in an organism is largely unchangeable. It is thus much more desirable to analyse the protein profiles of these cells, as proteins represent the current state of these cells much more accurate. Thus, they represent optimal biomarkers.

In research, proteomic mass spectrometry in particular is used as a method for the simultaneous detection of a large number of proteins. However, there are still major hurdles for drug development or, in the long term, for use in the clinic. In particular, there is a lack of standardised workflows for sample preparation and analysis of the smallest sample and protein quantities. The aim of the project KASPAR is to make protein profiles from single cells usable for drug development. For this purpose, a workflow consisting of automated sample preparation, AI-supported image analysis and data evaluation as well as mass spectrometry is being developed and validated using PDX.