AI for ATMP

AI-BASED QUALITY CONTROL METHODs FOR ATMP MANUFACTURING

Quality Control (QC) by manual assessment of morphology and assessment of typical biomarkers fails to depict subtle changes in cells that lead to lead to product failure. There is a need for new, cost-efficient characterization methods in production of cell based ATMPs. We will develop a platform for QC of human embryonic stem cell (hESC) starting material using an AI-based method and Deep Learning (DL), based on artificial neural networks (NN), to identify complex non-linear relationships present in global transcriptomic datasets. In order to make this cost-effective we will optimise the DL-based method to be used with gene expression data generated using single-cell quantitative PCR (sc-qPCR) technology.

Coordinator: University of Skövde

Partners: Takara Bio Europe AB, TATAA Biocenter AB, MultiD Analysis AB and RISE

WORK PACKAGES

  • Project management and dissemination
  • Single cell transcriptomics data generation
  • NN classifier development and interpretation
  • Transfer learning, final model validation and prototype generation
  • Business Development and Exploitation

Contact: Jane Synnergren