|ZFIN ID: ZDB-PUB-160907-2|
Real-time prediction of cell division timing in developing zebrafish embryo
Kozawa, S., Akanuma, T., Sato, T., Sato, Y.D., Ikeda, K., Sato, T.N.
|Source:||Scientific Reports 6: 32962 (Journal)|
|Registered Authors:||Akanuma, Takashi, Sato, Thomas N.|
|Keywords:||Bayesian inference, Time series|
|PubMed:||27597656 Full text @ Sci. Rep.|
Kozawa, S., Akanuma, T., Sato, T., Sato, Y.D., Ikeda, K., Sato, T.N. (2016) Real-time prediction of cell division timing in developing zebrafish embryo. Scientific Reports. 6:32962.
ABSTRACTCombination of live-imaging and live-manipulation of developing embryos in vivo provides a useful tool to study developmental processes. Identification and selection of target cells for an in vivo live-manipulation are generally performed by experience- and knowledge-based decision-making of the observer. Computer-assisted live-prediction method would be an additional approach to facilitate the identification and selection of the appropriate target cells. Herein we report such a method using developing zebrafish embryos. We choose V2 neural progenitor cells in developing zebrafish embryo as their successive shape changes can be visualized in real-time in vivo. We developed a relatively simple mathematical method of describing cellular geometry of V2 cells to predict cell division-timing based on their successively changing shapes in vivo. Using quantitatively measured 4D live-imaging data, features of V2 cell-shape at each time point prior to division were extracted and a statistical model capturing the successive changes of the V2 cell-shape was developed. By applying sequential Bayesian inference method to the model, we successfully predicted division-timing of randomly selected individual V2 cells while the cell behavior was being live-imaged. This system could assist pre-selecting target cells desirable for real-time manipulation-thus, presenting a new opportunity for in vivo experimental systems.