Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Comparative Study
. 2015 Feb 15;31(4):453-61.
doi: 10.1093/bioinformatics/btu407. Epub 2014 Jul 3.

Inter-species Prediction of Protein Phosphorylation in the Sbv IMPROVER Species Translation Challenge

Affiliations
Free PMC article
Comparative Study

Inter-species Prediction of Protein Phosphorylation in the Sbv IMPROVER Species Translation Challenge

Michael Biehl et al. Bioinformatics. .
Free PMC article

Abstract

Motivation: Animal models are widely used in biomedical research for reasons ranging from practical to ethical. An important issue is whether rodent models are predictive of human biology. This has been addressed recently in the framework of a series of challenges designed by the systems biology verification for Industrial Methodology for Process Verification in Research (sbv IMPROVER) initiative. In particular, one of the sub-challenges was devoted to the prediction of protein phosphorylation responses in human bronchial epithelial cells, exposed to a number of different chemical stimuli, given the responses in rat bronchial epithelial cells. Participating teams were asked to make inter-species predictions on the basis of available training examples, comprising transcriptomics and phosphoproteomics data.

Results: Here, the two best performing teams present their data-driven approaches and computational methods. In addition, post hoc analyses of the datasets and challenge results were performed by the participants and challenge organizers. The challenge outcome indicates that successful prediction of protein phosphorylation status in human based on rat phosphorylation levels is feasible. However, within the limitations of the computational tools used, the inclusion of gene expression data does not improve the prediction quality. The post hoc analysis of time-specific measurements sheds light on the signaling pathways in both species.

Availability and implementation: A detailed description of the dataset, challenge design and outcome is available at www.sbvimprover.com. The code used by team IGB is provided under http://github.com/uci-igb/improver2013. Implementations of the algorithms applied by team AMG are available at http://bhanot.biomaps.rutgers.edu/wiki/AMG-sc2-code.zip.

Contact: meikelbiehl@gmail.com.

Figures

Fig. 1.
Fig. 1.
Schematic illustration of the sub-challenge structure and datasets. The objective was to predict the phosphorylation status (P) of human phosphoproteins to stimuli subset B, shown in red, given the gene expression (GEx) and phosphorylation data for rat under the same stimuli. Available data (blue) also comprised the measurements of phosphorylation and gene expression in rat and human under a different set of stimuli A, which served as the training data. Human GEx data under the set of stimuli B was unavailable (shown in gray)
Fig. 2.
Fig. 2.
Color-coded visualization of the predictions for |humP|>3 in dataset B: cnaive (upper left panel), cLVQ (upper right) and the combination cAMG of team AMG (lower left). The lower right panel displays the prediction by team IGB. Proteins are numbered according to the list given in Section 2.1
Fig. 3.
Fig. 3.
Comparison of signaling pathways in rat and human. (A) Heatmap showing the clustering for directionality of phosphoprotein activation (columns) after a given stimulus (rows) for the two species, showing which phosphoproteins were activated early or late in each species by the different stimuli. Only stimuli with at least one non-zero entry according to Table 2 are shown. (B and C) Top: in orange are shown potential pathway activation diagrams for phosphoproteins activated by RPKB6S1 (B) and AKT1 (C). Bottom: left heatmaps show the clustering of the rat phosphorylation activation status of the phosphoproteins shown in the diagrams for all active stimuli. Right heatmaps display human phosphorylation activation status of the phosphoproteins shown in the diagrams for stimuli using the same clustering structure obtained from the rat data to ease comparison among species. Only stimuli where activation is present in at least one species are shown. Protein phosphorylation states are defined as inactive, active early (active only at 5 min), active at both time points (active at 5 and 25 min) and active late (active only at 25 min)

Similar articles

See all similar articles

Cited by 4 articles

References

    1. Baldi P, Sadowski P. Understanding dropout. Adv. Neural. Inf. Process. Syst. 2013;26:2814–2822.
    1. Baldi P, Sadowski P. The Dropout learning algorithm. Artificial Intelligence. 2014;210:78–122. - PMC - PubMed
    1. Bergstra J, et al. Proceedings of the Python for Scientific Computing Conference (SciPy) Austin, TX: 2010. Theano: a CPU and GPU math expression compiler.
    1. Biehl M, et al. Dynamics and generalization ability of LVQ algorithms. J. Mach. Learn. Res. 2007;8:323–360.
    1. Brodersen KH, et al. Proceedings of the 2010 20th International Conference on Pattern Recognition (ICPR’10) Washington, DC, USA: IEEE Computer Society; 2010. The balanced accuracy and its posterior distribution; pp. 3121–3124.

Publication types

Substances

Feedback