Limits...
Inferring the effective TOR-dependent network: a computational study in yeast.

Mohammadi S, Subramaniam S, Grama A - BMC Syst Biol (2013)

Bottom Line: These rankings must be normalized for degree bias, appropriately interpreted, and mapped to associated roles in pathways.We use this framework to identify the most relevant transcription factors in mediating the observed transcriptional response, and to construct the effective response network of the TOR pathway.It also allows us to identify potential network biomarkers for diagnosis and prognosis of age-related pathologies.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Computer Science, Purdue University, West Lafayette, Indiana, USA. mohammadi@purdue.edu.

ABSTRACT

Background: Calorie restriction (CR) is one of the most conserved non-genetic interventions that extends healthspan in evolutionarily distant species, ranging from yeast to mammals. The target of rapamycin (TOR) has been shown to play a key role in mediating healthspan extension in response to CR by integrating different signals that monitor nutrient-availability and orchestrating various components of cellular machinery in response. Both genetic and pharmacological interventions that inhibit the TOR pathway exhibit a similar phenotype, which is not further amplified by CR.

Results: In this paper, we present the first comprehensive, computationally derived map of TOR downstream effectors, with the objective of discovering key lifespan mediators, their crosstalk, and high-level organization. We adopt a systematic approach for tracing information flow from the TOR complex and use it to identify relevant signaling elements. By constructing a high-level functional map of TOR downstream effectors, we show that our approach is not only capable of recapturing previously known pathways, but also suggests potential targets for future studies.Information flow scores provide an aggregate ranking of relevance of proteins with respect to the TOR signaling pathway. These rankings must be normalized for degree bias, appropriately interpreted, and mapped to associated roles in pathways. We propose a novel statistical framework for integrating information flow scores, the set of differentially expressed genes in response to rapamycin treatment, and the transcriptional regulatory network. We use this framework to identify the most relevant transcription factors in mediating the observed transcriptional response, and to construct the effective response network of the TOR pathway. This network is hypothesized to mediate life-span extension in response to TOR inhibition.

Conclusions: Our approach, unlike experimental methods, is not limited to specific aspects of cellular response. Rather, it predicts transcriptional changes and post-translational modifications in response to TOR inhibition. The constructed effective response network greatly enhances understanding of the mechanisms underlying the aging process and helps in identifying new targets for further investigation of anti-aging regimes. It also allows us to identify potential network biomarkers for diagnosis and prognosis of age-related pathologies.

Show MeSH

Related in: MedlinePlus

Network integration process. Example of the network integration process around Sch9p. Protein-protein interactions (PPI) and post-translational modifications (PTM) were extracted from BioGRID dataset.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4016608&req=5

Figure 1: Network integration process. Example of the network integration process around Sch9p. Protein-protein interactions (PPI) and post-translational modifications (PTM) were extracted from BioGRID dataset.

Mentions: Given the yeast interactome, constructed using the procedure detailed in Methods Section and illustrated in Figure 1, we compute information flow scores using random walks initiated at selected nodes in the interactome. These nodes comprise members of the TORC1 complex, each of which propagates a unit flow (normalized to 0.2 for each of the five member proteins). We use a discrete random-walk process in which, at each step, every protein aggregates incoming signals and distributes them equally among outgoing neighbors. The final information flow scores are computed as the steady-state distribution of the random-walk process. One of the key parameters in the random-walk process, which controls the depth of propagation, is called the restart-probability. This is the probability that a random walker continues the walk (as opposed to teleporting to a node chosen from among a set of preferred nodes). In order to give all nodes in the interactome a chance of being visited, we use the relationship between restart probability and the mean depth of random-walks by setting parameter α to be equal to , where d is the diameter of the interactome. For the yeast interactome, we determine the diameter to be equal to 6 and set , correspondingly (please see the Methods section for details of information flow computations). Figure 2 illustrates the distribution of computed information flow scores, starting from TORC1, as a function of node distance from TORC1. It is evident from the figure that computed scores are functions of both distance from source nodes, as well as multiplicity of paths between source and sink nodes. This can be verified from the overlapping tails of distributions for nodes at different distances, as well as the varied distribution of scores among nodes at the same distance from TORC1. The final list of computed information flow scores is available for download as Additional file 1.


Inferring the effective TOR-dependent network: a computational study in yeast.

Mohammadi S, Subramaniam S, Grama A - BMC Syst Biol (2013)

Network integration process. Example of the network integration process around Sch9p. Protein-protein interactions (PPI) and post-translational modifications (PTM) were extracted from BioGRID dataset.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4016608&req=5

Figure 1: Network integration process. Example of the network integration process around Sch9p. Protein-protein interactions (PPI) and post-translational modifications (PTM) were extracted from BioGRID dataset.
Mentions: Given the yeast interactome, constructed using the procedure detailed in Methods Section and illustrated in Figure 1, we compute information flow scores using random walks initiated at selected nodes in the interactome. These nodes comprise members of the TORC1 complex, each of which propagates a unit flow (normalized to 0.2 for each of the five member proteins). We use a discrete random-walk process in which, at each step, every protein aggregates incoming signals and distributes them equally among outgoing neighbors. The final information flow scores are computed as the steady-state distribution of the random-walk process. One of the key parameters in the random-walk process, which controls the depth of propagation, is called the restart-probability. This is the probability that a random walker continues the walk (as opposed to teleporting to a node chosen from among a set of preferred nodes). In order to give all nodes in the interactome a chance of being visited, we use the relationship between restart probability and the mean depth of random-walks by setting parameter α to be equal to , where d is the diameter of the interactome. For the yeast interactome, we determine the diameter to be equal to 6 and set , correspondingly (please see the Methods section for details of information flow computations). Figure 2 illustrates the distribution of computed information flow scores, starting from TORC1, as a function of node distance from TORC1. It is evident from the figure that computed scores are functions of both distance from source nodes, as well as multiplicity of paths between source and sink nodes. This can be verified from the overlapping tails of distributions for nodes at different distances, as well as the varied distribution of scores among nodes at the same distance from TORC1. The final list of computed information flow scores is available for download as Additional file 1.

Bottom Line: These rankings must be normalized for degree bias, appropriately interpreted, and mapped to associated roles in pathways.We use this framework to identify the most relevant transcription factors in mediating the observed transcriptional response, and to construct the effective response network of the TOR pathway.It also allows us to identify potential network biomarkers for diagnosis and prognosis of age-related pathologies.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Computer Science, Purdue University, West Lafayette, Indiana, USA. mohammadi@purdue.edu.

ABSTRACT

Background: Calorie restriction (CR) is one of the most conserved non-genetic interventions that extends healthspan in evolutionarily distant species, ranging from yeast to mammals. The target of rapamycin (TOR) has been shown to play a key role in mediating healthspan extension in response to CR by integrating different signals that monitor nutrient-availability and orchestrating various components of cellular machinery in response. Both genetic and pharmacological interventions that inhibit the TOR pathway exhibit a similar phenotype, which is not further amplified by CR.

Results: In this paper, we present the first comprehensive, computationally derived map of TOR downstream effectors, with the objective of discovering key lifespan mediators, their crosstalk, and high-level organization. We adopt a systematic approach for tracing information flow from the TOR complex and use it to identify relevant signaling elements. By constructing a high-level functional map of TOR downstream effectors, we show that our approach is not only capable of recapturing previously known pathways, but also suggests potential targets for future studies.Information flow scores provide an aggregate ranking of relevance of proteins with respect to the TOR signaling pathway. These rankings must be normalized for degree bias, appropriately interpreted, and mapped to associated roles in pathways. We propose a novel statistical framework for integrating information flow scores, the set of differentially expressed genes in response to rapamycin treatment, and the transcriptional regulatory network. We use this framework to identify the most relevant transcription factors in mediating the observed transcriptional response, and to construct the effective response network of the TOR pathway. This network is hypothesized to mediate life-span extension in response to TOR inhibition.

Conclusions: Our approach, unlike experimental methods, is not limited to specific aspects of cellular response. Rather, it predicts transcriptional changes and post-translational modifications in response to TOR inhibition. The constructed effective response network greatly enhances understanding of the mechanisms underlying the aging process and helps in identifying new targets for further investigation of anti-aging regimes. It also allows us to identify potential network biomarkers for diagnosis and prognosis of age-related pathologies.

Show MeSH
Related in: MedlinePlus