Automated time-lapse 1009820-21-6 analysis can therefore identify perturbations that affect both interphase and mitosis, which would be misinterpreted if a single measurement of mitotic index were performed on fixed cells. Manual measurement of mitotic duration was performed by one experienced individual to ensure consistency. Images were analyzed in ImageJ . Mitotic duration was defined as the difference between the mitotic entry frame and the first frame showing chromatid segregation for each mitotic event visible throughout a movie. The frame of mitotic entry was defined as the first frame with nucleus showing early prophase characteristics, such as initiation of chromatin condensation. If an advanced prometaphase cell was identified as the first mitotic event, then the prior frame was chosen as the mitotic entry to avoid underestimating mitotic duration. Each entry and exit frame as well as the mitotic duration were recorded in excel files and the data was used to calculate the median and mean event duration and to plot the cumulative frequency curves, using the Tools/Data analysis/Histogram function of Excel, for each condition tested in an experiment. Interphase duration was determined to be the time difference between the first frame of segregation of the first mitosis to the mitotic entry frame of the following mitosis. Typically half of the movies analyzed automatically were analyzed manually, to limit time of analysis to a reasonable amount while ensuring good population representation. For comparison of differences between independent sets of measurements, the Mann-Whitney-Wilcoxon non-parametric test was performed. Lapatinib Significance level was alpha= 0.05. The Mann- Whitney-Wilcoxon test relies on comparing the ranks of sample measurements of two conditions rather than comparing the sample measurement values. The measurements of both samples are ranked together. Under the null hypothesis, the distribution of the two measurement samples is the same. Under the alternative hypothesis, there would be a difference in distribution of the two measurement samples, so measurements from one sample population would precede measurements from the second sample in the distribution. Nuclei were segmented from the background as described previously . The centroid for each nucleus was calculated in each frame. Starting from frame one, possible matches for each nucleus were identified within a 30 pixel radius. Best matches were identified based upon similarity of area, grey value histogram, XY displacement, speed, direction, shape and relationship with neighbors . Wavelet, geometric, moment, texture and shape descriptor features can be extracted from each nucleus identified for a total of 211 features for the SVM approach. Wang et al. fully reference the 211 features . The time series involved analysis of only 11 geometry features and Area and Intensity were the only features necessary for identifying the transitions between interphase and mitosis.