Periodic Reporting for period 1 - AiPSC (AI-powered platform for autologous iPSC manufacturing)
Période du rapport: 2022-09-01 au 2023-08-31
The AiPSC seeks to develop a new technology that will enable the mass production of personalized iPSC-based therapies. The consortium will for the first time create an artificial intelligence (AI) guided microfluidic device that standardizes the GMP production of autologous iPSCs fast and at a fraction of the current cost. Moreover, it will conduct cutting-edge cell genomics and bioinformatics research on iPSCs to identify clones of the highest quality and develop a database that will be the basis for AI software to select clones that meet clinical standards. The consortium comprises experts in microfluidics engineering process automation for cell therapies, stem cell molecular biology and bioinformatics, GMP production, and AI modeling. Altogether, we propose to create revolutionary technology for low-cost, fast, and standardized automated mass production of autologous iPSCs, which holds the potential to enable numerous new therapies and make them accessible to the public.
References:
1 Takahashi, K. et al. Cell 126, 663–676 (2006)
2 Chinen, J. et al. J. Allergy Clin. Immunol. 125, S324 (2010)
3 Caldwell, K. J. et al. Front. Immunol. 11, 618427 (2021)
4 Zhao, W. et al. (2020)
5 Caldwell, A. et al. (2018)
6 Rossi, S. J. et al. Drug Saf. 9, 104–131 (1993)
MIDA established the basis for the generation of iPSCs in microfluidic devices:
-Developed and validated GMP-translated protocols for the generation and maintenance of autologous iPSCs.
-Developed a plan for iPSC quality assessment.
-Successfully downscaled the fibroblast reprogramming protocols to suit low-volume culture vessels such as microfluidic chips.
-Created an annotated dataset of more than 4000 reprogramming images for training the AI models for iPSC colony detection and selection.
KVentures in order to initiate the development of the AI algorithms for the software designed for determining iPSC colony quality has:
-Developed an annotation platform for securely depositing and annotating reprogramming datasets.
-Develop and train supervised and unsupervised AI models for iPSC colony detection and the morphological selection of high-quality colonies
-Introduced a scheduler command line for enabling time-lapse imaging of reprogramming experiments
-Established two distinct stitching approaches for the analysis of whole slide images.
Leiden University in order to expand and refine the quality standards for iPSCs, will during the course of the project create a new colony quality index based on next-generation sequencing and ATAC-seq. During the reporting period, they have:
-Curated an extensive compilation of next-generation sequencing datasets obtained from iPSCs generated through various methodologies, utilizing diverse cell sources and donors, including primary cells and cancer cell lines.
-Acquired cell lines from the same donor originating from different cell sources such as blood, urine, and skin.
-Devised comprehensive protocols for differentiating these cells from the same donor and plans to employ single-cell RNA and ATAC sequencing techniques to extract valuable insights into the differentiation potential of iPSCs.
MIRCOD has developed the first designs of microfluidic automation for iPSC production:
-Designed the first AiPSC reprogramming microfluidic chips
-Designed the automatic liquid handling of the microfluidic chips