Figure - PMC

Skip to main content

An official website of the United States government

Here's how you know

Here's how you know

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

View full-text article in PMC

. 2012 Aug 24;28(21):2804–2810. doi: 10.1093/bioinformatics/bts514

Search in PMC
Search in PubMed
View in NLM Catalog
Add to search

© The Author 2012. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com

PMC Copyright notice

Fig. 3. — Data-driven signaling topology in the breast cancer cell line MDA-MB-468. Reverse-phase protein arrays were used to interrogate the phospho-proteins shown in (a), including key components of AKT, MAPK and STAT pathways, through time. DBNs were used to integrate the data with a network prior, derived from existing biology (Supplementary Fig. S3). Prior strength was set objectively using an empirical Bayes approach, resulting in (see text and Supplementary Fig. S5). (a) Edges represent probabilistic relationships between proteins, through time. Edge labels indicate corresponding (posterior) probabilities (calculated exactly; edge thickness proportional to probability; all edges with probability 0.4 shown; strikethroughs ‘/’ indicate edges not expected under the network prior; Supplementary Figure S6 shows heatmap of all 400 posterior edge probabilities; list of proteins (and antibodies) given in Supplementary Table S1). (b) Sensitivity to prior strength. Results were compared over a range of values of plus the flat prior (i.e. ) and prior only. Heatmap shows Pearson correlations between edge probabilities for pairs of prior regimes. Results were not sensitive to precise value of and differed markedly from prior alone. (c) Sensitivity to prior graph. The prior graph was perturbed and results obtained compared with those reported (see Supplementary Text for details). Correlation (as in (b)) is shown as a function of number of edge changes in the prior graph (‘Structural Hamming Distance’, see Supplementary Fig. S7 for error bars)

Inline graphic — Data-driven signaling topology in the breast cancer cell line MDA-MB-468. Reverse-phase protein arrays were used to interrogate the phospho-proteins shown in (a), including key components of AKT, MAPK and STAT pathways, through time. DBNs were used to integrate the data with a network prior, derived from existing biology (Supplementary Fig. S3). Prior strength was set objectively using an empirical Bayes approach, resulting in (see text and Supplementary Fig. S5). (a) Edges represent probabilistic relationships between proteins, through time. Edge labels indicate corresponding (posterior) probabilities (calculated exactly; edge thickness proportional to probability; all edges with probability 0.4 shown; strikethroughs ‘/’ indicate edges not expected under the network prior; Supplementary Figure S6 shows heatmap of all 400 posterior edge probabilities; list of proteins (and antibodies) given in Supplementary Table S1). (b) Sensitivity to prior strength. Results were compared over a range of values of plus the flat prior (i.e. ) and prior only. Heatmap shows Pearson correlations between edge probabilities for pairs of prior regimes. Results were not sensitive to precise value of and differed markedly from prior alone. (c) Sensitivity to prior graph. The prior graph was perturbed and results obtained compared with those reported (see Supplementary Text for details). Correlation (as in (b)) is shown as a function of number of edge changes in the prior graph (‘Structural Hamming Distance’, see Supplementary Fig. S7 for error bars)