Figure 2
Computer vision based detection of human urinal cells (UCs) derived iPSC colonies. (a) Example of training patches. Positive samples are parts of iPSC colonies, and negative sample are other cells or background. Scale bar, \(200\mu {\rm{m}}\). (b) A detection map. Example of iPSC colonies detection in one 6-well plate at Day 21 after reprogramming induction. Red regions indicate the position of iPSC colonies, which were detected by bright field image classification algorithm. A small area is enlarged to show the details of the detection. (c) Characterization of iPSCs, which were manually picked guided by the binary map. iPSC colonies (n = 3) were picked into cell culture plate based on the binary map. After 4–7 days culturing, iPSCs were stained with antibodies against known pluripotency surface markers (left). The nucleus was stained by DAPI (middle), merged images of two channel (right). Scale bar, 50μm. (d) Selected iPSCs were analysed for Nanog, SOX2 and OCT4 expression by qRT-PCR. Human ES cell line H1 was used as positive control. Error bars indicate  ± s.d. of triplicates. P value is referenced to UCs. **Indicates P < 0.01. (e) PCR detection of exogenous eipsomal DNA in UC-iPSC. UC(Passage 3) and H1(Passage 46) cell were severed as the negative control, UC-iPSC(Passage21) was stably expanded, UC transiently transfected with eipsomal vector (pEP4EO2SET2K and pCEP4- miR-302-367 cluster) was served as the positive control.
