Table 1.
Model comparison for the number of mixtures component. The table shows the Bayes factors for the number of Gaussian components needed to model the distribution of Nanog-GFP data in various culture conditions: B32 is the odds of three components over two, B34 the odds of three components over four and B42 the odds of four components over two. Odds between 1 and 3 are marginal, between 3 and 10 substantial and above 10 strong.
| LIF+BMP | Chiron | PD03 | 2i | |
|---|---|---|---|---|
| B32 | 113 | 713 | 39.2 | 1/38560 |
| B34 | 2.03 | 2.33 | 2.14 | 1/6247 |
| B42 | 55.7 | 306 | 18.3 | 1/6.17 |