Limits...
Inferring interaction type in gene regulatory networks using co-expression data.

Khosravi P, Gazestani VH, Pirhaji L, Law B, Sadeghi M, Goliaei B, Bader GD - Algorithms Mol Biol (2015)

Bottom Line: Currently information-based approaches are widely used for inferring gene regulatory interactions from genomics data, such as gene expression profiles; however, these approaches do not provide evidence about the regulation type (positive or negative sign) of the interaction.Our new method has approximately 68, 70, and 100 percent accuracy, respectively, for these networks.It can serve as a complementary approach for a wide range of network reconstruction methods that do not provide information about the interaction type.

View Article: PubMed Central - PubMed

Affiliation: School of Biological Sciences, Institute for Research in Fundamental Sciences (IPM), Tehran, Iran ; The Donnelly Centre, University of Toronto, Toronto, Canada.

ABSTRACT

Background: Knowledge of interaction types in biological networks is important for understanding the functional organization of the cell. Currently information-based approaches are widely used for inferring gene regulatory interactions from genomics data, such as gene expression profiles; however, these approaches do not provide evidence about the regulation type (positive or negative sign) of the interaction.

Results: This paper describes a novel algorithm, "Signing of Regulatory Networks" (SIREN), which can infer the regulatory type of interactions in a known gene regulatory network (GRN) given corresponding genome-wide gene expression data. To assess our new approach, we applied it to three different benchmark gene regulatory networks, including Escherichia coli, prostate cancer, and an in silico constructed network. Our new method has approximately 68, 70, and 100 percent accuracy, respectively, for these networks. To showcase the utility of SIREN algorithm, we used it to predict previously unknown regulation types for 454 interactions related to the prostate cancer GRN.

Conclusions: SIREN is an efficient algorithm with low computational complexity; hence, it is applicable to large biological networks. It can serve as a complementary approach for a wide range of network reconstruction methods that do not provide information about the interaction type.

No MeSH data available.


Related in: MedlinePlus

Deciphering interaction type from co-expression patterns. a, b The two-dimensional grids are constructed for two selected gene pairs with known activating or inhibitory effect on each other. The color density for each cell in the grid represents the computed PMI for that cell multiplied in the occurrence probability of the cell and corresponding rescaling value (determined based on the rescaling matrix). The PMI and occurrence probability is calculated based on the associated transcriptome data. Red indicates positive score and blue represents negative score (defined based on the M3 rescaling matrix). SIREN score is determined by summing up the calculated values for each combination of bins. a For two example genes with known activating relationships (JUN and ATF3), cells defined as activating have non-zero values and cells defined as inhibitory relationship, have zero values. b This situation is reversed for two genes with known inhibitory relationship (ZEB1 and CDH1).
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4495944&req=5

Fig9: Deciphering interaction type from co-expression patterns. a, b The two-dimensional grids are constructed for two selected gene pairs with known activating or inhibitory effect on each other. The color density for each cell in the grid represents the computed PMI for that cell multiplied in the occurrence probability of the cell and corresponding rescaling value (determined based on the rescaling matrix). The PMI and occurrence probability is calculated based on the associated transcriptome data. Red indicates positive score and blue represents negative score (defined based on the M3 rescaling matrix). SIREN score is determined by summing up the calculated values for each combination of bins. a For two example genes with known activating relationships (JUN and ATF3), cells defined as activating have non-zero values and cells defined as inhibitory relationship, have zero values. b This situation is reversed for two genes with known inhibitory relationship (ZEB1 and CDH1).

Mentions: Figure 9 demonstrates the application of SIREN (with optimized parameters) on two example interactions, one activating and one inhibitory. As mentioned earlier, SIREN detects the interaction type in four steps: (1) discretizing the expression profile of each gene into 10 bins; (2) calculating the co-occurrence probability for each combination of bins using the PMI metric; (3) defining an activating or inhibitory relationship for each combination of bins by aid of a rescaling matrix; and (4) calculating a final score by integrating the calculated values for each combination of bins. For example, with two genes with a known activating relationship from STRING (JUN and ATF3) [52, 53], bins defined as activating have non-zero values while bins defined as inhibitory have mostly zero values (Figure 9a). This situation is reversed for two genes with a known inhibitory relationship (ZEB1 and CDH1) [54] (Figure 9b).Figure 9


Inferring interaction type in gene regulatory networks using co-expression data.

Khosravi P, Gazestani VH, Pirhaji L, Law B, Sadeghi M, Goliaei B, Bader GD - Algorithms Mol Biol (2015)

Deciphering interaction type from co-expression patterns. a, b The two-dimensional grids are constructed for two selected gene pairs with known activating or inhibitory effect on each other. The color density for each cell in the grid represents the computed PMI for that cell multiplied in the occurrence probability of the cell and corresponding rescaling value (determined based on the rescaling matrix). The PMI and occurrence probability is calculated based on the associated transcriptome data. Red indicates positive score and blue represents negative score (defined based on the M3 rescaling matrix). SIREN score is determined by summing up the calculated values for each combination of bins. a For two example genes with known activating relationships (JUN and ATF3), cells defined as activating have non-zero values and cells defined as inhibitory relationship, have zero values. b This situation is reversed for two genes with known inhibitory relationship (ZEB1 and CDH1).
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4495944&req=5

Fig9: Deciphering interaction type from co-expression patterns. a, b The two-dimensional grids are constructed for two selected gene pairs with known activating or inhibitory effect on each other. The color density for each cell in the grid represents the computed PMI for that cell multiplied in the occurrence probability of the cell and corresponding rescaling value (determined based on the rescaling matrix). The PMI and occurrence probability is calculated based on the associated transcriptome data. Red indicates positive score and blue represents negative score (defined based on the M3 rescaling matrix). SIREN score is determined by summing up the calculated values for each combination of bins. a For two example genes with known activating relationships (JUN and ATF3), cells defined as activating have non-zero values and cells defined as inhibitory relationship, have zero values. b This situation is reversed for two genes with known inhibitory relationship (ZEB1 and CDH1).
Mentions: Figure 9 demonstrates the application of SIREN (with optimized parameters) on two example interactions, one activating and one inhibitory. As mentioned earlier, SIREN detects the interaction type in four steps: (1) discretizing the expression profile of each gene into 10 bins; (2) calculating the co-occurrence probability for each combination of bins using the PMI metric; (3) defining an activating or inhibitory relationship for each combination of bins by aid of a rescaling matrix; and (4) calculating a final score by integrating the calculated values for each combination of bins. For example, with two genes with a known activating relationship from STRING (JUN and ATF3) [52, 53], bins defined as activating have non-zero values while bins defined as inhibitory have mostly zero values (Figure 9a). This situation is reversed for two genes with a known inhibitory relationship (ZEB1 and CDH1) [54] (Figure 9b).Figure 9

Bottom Line: Currently information-based approaches are widely used for inferring gene regulatory interactions from genomics data, such as gene expression profiles; however, these approaches do not provide evidence about the regulation type (positive or negative sign) of the interaction.Our new method has approximately 68, 70, and 100 percent accuracy, respectively, for these networks.It can serve as a complementary approach for a wide range of network reconstruction methods that do not provide information about the interaction type.

View Article: PubMed Central - PubMed

Affiliation: School of Biological Sciences, Institute for Research in Fundamental Sciences (IPM), Tehran, Iran ; The Donnelly Centre, University of Toronto, Toronto, Canada.

ABSTRACT

Background: Knowledge of interaction types in biological networks is important for understanding the functional organization of the cell. Currently information-based approaches are widely used for inferring gene regulatory interactions from genomics data, such as gene expression profiles; however, these approaches do not provide evidence about the regulation type (positive or negative sign) of the interaction.

Results: This paper describes a novel algorithm, "Signing of Regulatory Networks" (SIREN), which can infer the regulatory type of interactions in a known gene regulatory network (GRN) given corresponding genome-wide gene expression data. To assess our new approach, we applied it to three different benchmark gene regulatory networks, including Escherichia coli, prostate cancer, and an in silico constructed network. Our new method has approximately 68, 70, and 100 percent accuracy, respectively, for these networks. To showcase the utility of SIREN algorithm, we used it to predict previously unknown regulation types for 454 interactions related to the prostate cancer GRN.

Conclusions: SIREN is an efficient algorithm with low computational complexity; hence, it is applicable to large biological networks. It can serve as a complementary approach for a wide range of network reconstruction methods that do not provide information about the interaction type.

No MeSH data available.


Related in: MedlinePlus