School of Science Division of Life Science 34 CRISPR/Cas9 Analysis of Essential Genes Supervisor: POON Randy Yat Choi / LIFS Student: ZENG Qianhui / BCB Course: UROP 1100, Summer My study in this summer focused on two cell lines, HtTA1 and A9, which are all derived from Hela cells and stably transfected with tTA, a tetracycline-controlled translation activator. What is different between the two cell lines is that the expression of endogenous Cyclin A in A9 is completely silenced, and it is transfected with a specific DNA fragment that allows it to express cyclin A under the restriction by doxycycline. The experiment of serum starvation and the addition of doxycycline and IAA, adopting several techniques including cell culture, LCI and Western blot, proved their essential effect on cell growth and cyclin A protein expression respectively. Development of an Algorithm to Analyse Live-Cell Imaging Data Supervisor: POON Randy Yat Choi / LIFS Student: LEE Geunhee / BIOT Course: UROP 1100, Summer Dihydrocytochalasin B (DCB) inhibits actin polymerization critical for the assembly of the cell’s contractile ring during cytokinesis, leading to cytokinesis failure and tetraploidization. Treating U2OS cells with 10 μM of DCB induced tetraploidy. Live-cell imaging using fluorescence microscopy was performed for 24 hours to track the progression of individual tetraploid cells through subsequent mitosis. Mitotic events were manually recorded and categorized in cells expressing histone H2B-GFP. My results revealed that DCB-induced binucleated cells re-entered mitosis. About 50% of these cells successfully underwent pseudo-bipolar division, with four centrosomes segregated into two each in the two daughter cells. About 18% displayed asymmetric centrosome segregation (3 vs. 1), while about 32% of cells underwent multipolar division. Cell tracking is a computational method that automatically monitors individual cells over time in time-lapse microscopy videos. It provides a quantitative framework to follow cellular behavior and dynamic processes at the single-cell level. According to the report on the deep learning algorithm Cellpose, human operators can segment only 300-600 objects per hour, whereas fully automated methods provide advantages such as reduced human effort, increased reproducibility, and better scalability for large datasets. By using automated process, the consuming time can be reduced tremendously. Using a fine-tuned model of Cellpose, I showed that the accuracy of detecting the cells can be increased compared to using traditional segmentation programs such as LoG detector.
RkJQdWJsaXNoZXIy NDk5Njg=