Display Settings:

Items per page

Results: 6

1.
Figure 6

Figure 6. Non-Inhibition Identification.. From: Network Class Superposition Analyses.

The ROC for identifying non-interactions between components in the yeast cell network.

Carl A. B. Pearson, et al. PLoS One. 2013;8(4):e59046.
2.
Figure 5

Figure 5. Activation Identification.. From: Network Class Superposition Analyses.

The ROC for identifying the flexible and free activation interactions in the yeast cell network.

Carl A. B. Pearson, et al. PLoS One. 2013;8(4):e59046.
3.
Figure 4

Figure 4. Inhibition Identification.. From: Network Class Superposition Analyses.

The ROC for identifying the flexible and free inhibition interactions in the yeast cell network.

Carl A. B. Pearson, et al. PLoS One. 2013;8(4):e59046.
4.
Figure 3

Figure 3. Experimental Selection.. From: Network Class Superposition Analyses.

This plot is a histogram comparing number of experiments to determine an underlying network given partial initial information. The categories are the number of simulated experiments using -derived entropy () minus the number using using naive entropy (), so the black region corresponds to the relative amount where performs better, and the leftmost columns correspond to the greatest advantage. This sample mixes the results from considering 10 k supporting networks from each of our 11-element random dynamics. Similar results were obtained for other system sizes.

Carl A. B. Pearson, et al. PLoS One. 2013;8(4):e59046.
5.
Figure 1

Figure 1. Point Attractor Distribution Correlations.. From: Network Class Superposition Analyses.

This plot compares the distribution of point attractor number as calculated based on versus sampled. Each point, with base-10 logarithmic scales, is , the calculated probability () of some number () of point attractors based on from a particular dynamic , and , the sampled frequency () of that number of point attractors in networks that support ; or, succinctly: . The plot shows that the -based distribution is tightly correlated with the sample down to about the level, and then skews low. The plot is annotated with other information: both axes include rug plots to indicate point density, red horizontal lines indicating the 1 to 25 counts out the sample size (the region with visible notable “lines” of sample data), a blue 1-to-1 line for exact correlation, and dashed grey lines indicating the 20%, median, and 80% slices in the spread.

Carl A. B. Pearson, et al. PLoS One. 2013;8(4):e59046.
6.
Figure 2

Figure 2. Derrida Plot Comparing Yeast Cell Network and Cell Cycle Process T.. From: Network Class Superposition Analyses.

This plot compares the putative yeast cell cycle network (grey) and supporting the yeast cell cycle process (black). We have extended the plot to show median (solid points) plus box-and-whiskers in addition to means (open points). For this size system, we were able to evaluate a complete population instead of sampling. The putative cell cycle network is stable according to the Derrida coefficient: i.e., the mean values form a trajectory below the line. The -based results require more interpretation and context. Notably, states with - i.e., identical states - have - i.e., the same outcome - on a set network with deterministic rules. If we use as a surrogate for such a network, then the resulting spread in for - essentially half the available range - should give pause when comparing the outcomes of nearby initial conditions. However, seems plausibly useful for estimating divergence for disparate initial conditions. On the other hand, if we believe our system is well represented by the superposition of networks, then that low spread in may provide insight into how (un)constrained the system noise is by the structured component.

Carl A. B. Pearson, et al. PLoS One. 2013;8(4):e59046.

Display Settings:

Items per page

Supplemental Content

Recent activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...
Write to the Help Desk