Limits...
An integrative analysis of meningioma tumors reveals the determinant genes and pathways of malignant transformation.

Iglesias Gómez JC, Mosquera Orgueira A - Front Oncol (2014)

Bottom Line: Thus, this study is aimed to identify the genomic and transcriptomic factors influencing evolution from benignity toward aggressive phenotypes.By applying an integrative bioinformatics pipeline combining public information on a wealth of biological layers of complexity (from genetic polymorphisms to protein interactions), this study identified a module of co-expressed genes highly correlated with tumor stage and statistically linked to several genomic regions (module Quantitative Trait Loci, mQTLs).As a result, cytoskeleton and cell-cell adhesion pathways, calcium-channels and glutamate receptors, as well as oxidoreductase and endoplasmic reticulum-associated degradation pathways were found to be the most important and redundant findings associated to meningioma progression.

View Article: PubMed Central - PubMed

Affiliation: Independent Video Editor , Santiago de Compostela , Spain.

ABSTRACT
Meningiomas are frequent central nervous system neoplasms, which despite their predominant benignity, show sporadically malignant behavior. Type 2 neurofibromatosis and polymorphisms in several genes have been associated with meningioma risk and are probably involved in its pathogenesis. Although GWAS studies have found loci related to meningioma risk, little is known about the factors determining malignant transformation. Thus, this study is aimed to identify the genomic and transcriptomic factors influencing evolution from benignity toward aggressive phenotypes. By applying an integrative bioinformatics pipeline combining public information on a wealth of biological layers of complexity (from genetic polymorphisms to protein interactions), this study identified a module of co-expressed genes highly correlated with tumor stage and statistically linked to several genomic regions (module Quantitative Trait Loci, mQTLs). Ontology analysis of the transcription hub genes identified microtubule-associated cell-cycle processes as key drivers of such network. mQTLs and single nucleotide polymorphisms associated with meningioma stage were replicated in an alternative meningioma cohort, and integration of these results with up-to-date scientific literature and several databases retrieved a list of genes and pathways with a potentially important role in meningioma malignancy. As a result, cytoskeleton and cell-cell adhesion pathways, calcium-channels and glutamate receptors, as well as oxidoreductase and endoplasmic reticulum-associated degradation pathways were found to be the most important and redundant findings associated to meningioma progression. This study presents an integrated view of the pathways involved in meningioma malignant conversion and paves the way for the development of new research lines that will improve our understanding of meningioma biology.

No MeSH data available.


Related in: MedlinePlus

mQTL plot. The colored upper part of the plot indicates the association of each SNP with meningioma stages (0: meningioma WHO Grade I, 1: meningioma WHO Grade II and III). The significance value is expressed as −log (P-value). The lower, blue part of the plot indicates the strength of the association of each SNP and the module.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4066933&req=5

Figure 5: mQTL plot. The colored upper part of the plot indicates the association of each SNP with meningioma stages (0: meningioma WHO Grade I, 1: meningioma WHO Grade II and III). The significance value is expressed as −log (P-value). The lower, blue part of the plot indicates the strength of the association of each SNP and the module.

Mentions: Weighted gene co-expression network analysis functions were applied to expression data according to several online tutorials (18). The adjacency matrix was calculated using a soft-thresholding power of 6, which showed an approximate scale-free topology (R2 = 0.75). Briefly, the connectivity value of each selected transcript (calculated similarly to kIN in Section “Regression Models”) was used to create a group of 10 bins with equal size, and each connectivity value was assigned to each bin. The connectivity distribution k was defined as the average connectivity value for each bin, whilst the probability distribution of k p(k) was defined as the ratio of the number of connectivity values in each bin by the number of connectivity values studied. Under the approximate scale-free topology assumption, the logarithm of p(k) (log(p(k))) and the logarithm of k (log(k)) are strongly negatively correlated. In this case, by using a soft-thresholding power of 6 we obtained a R2 = 0.75 (Pearson’s correlation of 0.86, regression slope of −1.46), whilst using no thresholding at all we obtained R2 = 0.03 (Pearson’s correlation of 0.17, regression slope of 1.05). Thus, the soft-thresholding value selected ensures approximated scale-free topology whilst retaining a higher number of informative connections in the network. Several co-expression modules were determined, and correlation between phenotypic data and their respective first principal components (a.k.a. module eigengenes, ME) was calculated. Mean gene significance with WHO Grade was calculated as the absolute average correlation of all module genes with this trait. Due to its marked positive correlation with several parameters (refer to Figure 5), the pink-module was chosen for further analysis. Module membership (MM, a.k.a kME) was defined as the Spearman’s correlation between the ME and the genes corresponding to the pink-module, which is considered a measure of centrality of each gene in the network. Gene-expression standard deviation was determined with the function rowSds, part of the package matrixStats (25).


An integrative analysis of meningioma tumors reveals the determinant genes and pathways of malignant transformation.

Iglesias Gómez JC, Mosquera Orgueira A - Front Oncol (2014)

mQTL plot. The colored upper part of the plot indicates the association of each SNP with meningioma stages (0: meningioma WHO Grade I, 1: meningioma WHO Grade II and III). The significance value is expressed as −log (P-value). The lower, blue part of the plot indicates the strength of the association of each SNP and the module.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4066933&req=5

Figure 5: mQTL plot. The colored upper part of the plot indicates the association of each SNP with meningioma stages (0: meningioma WHO Grade I, 1: meningioma WHO Grade II and III). The significance value is expressed as −log (P-value). The lower, blue part of the plot indicates the strength of the association of each SNP and the module.
Mentions: Weighted gene co-expression network analysis functions were applied to expression data according to several online tutorials (18). The adjacency matrix was calculated using a soft-thresholding power of 6, which showed an approximate scale-free topology (R2 = 0.75). Briefly, the connectivity value of each selected transcript (calculated similarly to kIN in Section “Regression Models”) was used to create a group of 10 bins with equal size, and each connectivity value was assigned to each bin. The connectivity distribution k was defined as the average connectivity value for each bin, whilst the probability distribution of k p(k) was defined as the ratio of the number of connectivity values in each bin by the number of connectivity values studied. Under the approximate scale-free topology assumption, the logarithm of p(k) (log(p(k))) and the logarithm of k (log(k)) are strongly negatively correlated. In this case, by using a soft-thresholding power of 6 we obtained a R2 = 0.75 (Pearson’s correlation of 0.86, regression slope of −1.46), whilst using no thresholding at all we obtained R2 = 0.03 (Pearson’s correlation of 0.17, regression slope of 1.05). Thus, the soft-thresholding value selected ensures approximated scale-free topology whilst retaining a higher number of informative connections in the network. Several co-expression modules were determined, and correlation between phenotypic data and their respective first principal components (a.k.a. module eigengenes, ME) was calculated. Mean gene significance with WHO Grade was calculated as the absolute average correlation of all module genes with this trait. Due to its marked positive correlation with several parameters (refer to Figure 5), the pink-module was chosen for further analysis. Module membership (MM, a.k.a kME) was defined as the Spearman’s correlation between the ME and the genes corresponding to the pink-module, which is considered a measure of centrality of each gene in the network. Gene-expression standard deviation was determined with the function rowSds, part of the package matrixStats (25).

Bottom Line: Thus, this study is aimed to identify the genomic and transcriptomic factors influencing evolution from benignity toward aggressive phenotypes.By applying an integrative bioinformatics pipeline combining public information on a wealth of biological layers of complexity (from genetic polymorphisms to protein interactions), this study identified a module of co-expressed genes highly correlated with tumor stage and statistically linked to several genomic regions (module Quantitative Trait Loci, mQTLs).As a result, cytoskeleton and cell-cell adhesion pathways, calcium-channels and glutamate receptors, as well as oxidoreductase and endoplasmic reticulum-associated degradation pathways were found to be the most important and redundant findings associated to meningioma progression.

View Article: PubMed Central - PubMed

Affiliation: Independent Video Editor , Santiago de Compostela , Spain.

ABSTRACT
Meningiomas are frequent central nervous system neoplasms, which despite their predominant benignity, show sporadically malignant behavior. Type 2 neurofibromatosis and polymorphisms in several genes have been associated with meningioma risk and are probably involved in its pathogenesis. Although GWAS studies have found loci related to meningioma risk, little is known about the factors determining malignant transformation. Thus, this study is aimed to identify the genomic and transcriptomic factors influencing evolution from benignity toward aggressive phenotypes. By applying an integrative bioinformatics pipeline combining public information on a wealth of biological layers of complexity (from genetic polymorphisms to protein interactions), this study identified a module of co-expressed genes highly correlated with tumor stage and statistically linked to several genomic regions (module Quantitative Trait Loci, mQTLs). Ontology analysis of the transcription hub genes identified microtubule-associated cell-cycle processes as key drivers of such network. mQTLs and single nucleotide polymorphisms associated with meningioma stage were replicated in an alternative meningioma cohort, and integration of these results with up-to-date scientific literature and several databases retrieved a list of genes and pathways with a potentially important role in meningioma malignancy. As a result, cytoskeleton and cell-cell adhesion pathways, calcium-channels and glutamate receptors, as well as oxidoreductase and endoplasmic reticulum-associated degradation pathways were found to be the most important and redundant findings associated to meningioma progression. This study presents an integrated view of the pathways involved in meningioma malignant conversion and paves the way for the development of new research lines that will improve our understanding of meningioma biology.

No MeSH data available.


Related in: MedlinePlus