However, how the cell state alteration process happens from terminal differentiation to pluripotency is unclear. The similarities and differences in the transcriptomes of iPSCs and ES cells have been estimated, while other properties of iPSCs are also different compared with ESCs, such as the genome methylation state, microRNA profiling, histone modification, proteomic profiles, and so on. It is still a challenge to find an accurate and easy method to estimate the pluripotency of iPSC candidates based on these cellular properties. The value of iPSCs is their pluripotency. From this perspective, pluripotency should be a gold standard for estimating the quality of iPSCs. The tetraploid complementation assay, with is the most strict standard, has been successfully performed on mouse cells.

Our results suggested that genome-wide expression patterns could partly reflect the pluripotency of mouse cells. The Distance index of dataset GSE16925 indicated that low quality iPS cells distinctly have bigger Di then the high quality iPS cells, and this disparity is also clearly reflected by the success of live pups. We believe that the Distance index, as a more accurate and reasonable measurement, have the potential to become an easy standard to estimate the quality of iPSCs at molecular level. The similarity defined by hierarchical clustering method severely depends on the mathematical characteristics of expression profiles. The system error of ES cells expression profiles would affect the clustering results. In our model, the “developing lines” generated by time-ordered linear model have distinct biological meaning: such lines are projection of ES cells differentiation trajectories. Meanwhile, the calculation of Distance index by this supervised method is based on a large number of expression profiles from different laboratories, and these existing datasets give our method greater robustness and accuracy.

Such characteristics enable us to compare expression profiles of different sources more easily. The dynamic changes in cell states induced by reprogramming were also clearly indicated by the “Differentiation-index coordinate”. These dynamic changes of cell states would help us to understand more about the movement trajectories of the ES cells differentiation and the reprogramming process of somatic cells. As shown, the time-order linear model was also a novel method to analyze time-resolved experimental data. This method generated lists of the significantly up/down regulated genes during the timeresolved experiment. Based on the Protein-protein interaction network and significantly changed genes during human ES cell differentiation, we identified some interesting “seesaw”modules.