Limits...
Shared Pathways Among Autism Candidate Genes Determined by Co-expression Network Analysis of the Developing Human Brain Transcriptome.

Mahfouz A, Ziats MN, Rennert OM, Lelieveldt BP, Reinders MJ - J. Mol. Neurosci. (2015)

Bottom Line: Autism spectrum disorder (ASD) is a neurodevelopmental syndrome known to have a significant but complex genetic etiology.Furthermore, we also constructed co-expression networks from the entire transcriptome and found that ASD candidate genes were enriched in modules related to mitochondrial function, protein translation, and ubiquitination.Overall, our multi-dimensional co-expression analysis of ASD candidate genes in the normal developing human brain suggests the heterogeneous set of ASD candidates share transcriptional networks related to synapse formation and elimination, protein turnover, and mitochondrial function.

View Article: PubMed Central - PubMed

Affiliation: Delft Bioinformatics Lab, Delft University of Technology, Delft, The Netherlands. a.mahfouz@tudelft.nl.

ABSTRACT
Autism spectrum disorder (ASD) is a neurodevelopmental syndrome known to have a significant but complex genetic etiology. Hundreds of diverse genes have been implicated in ASD; yet understanding how many genes, each with disparate function, can all be linked to a single clinical phenotype remains unclear. We hypothesized that understanding functional relationships between autism candidate genes during normal human brain development may provide convergent mechanistic insight into the genetic heterogeneity of ASD. We analyzed the co-expression relationships of 455 genes previously implicated in autism using the BrainSpan human transcriptome database, across 16 anatomical brain regions spanning prenatal life through adulthood. We discovered modules of ASD candidate genes with biologically relevant temporal co-expression dynamics, which were enriched for functional ontologies related to synaptogenesis, apoptosis, and GABA-ergic neurons. Furthermore, we also constructed co-expression networks from the entire transcriptome and found that ASD candidate genes were enriched in modules related to mitochondrial function, protein translation, and ubiquitination. Hub genes central to these ASD-enriched modules were further identified, and their functions supported these ontological findings. Overall, our multi-dimensional co-expression analysis of ASD candidate genes in the normal developing human brain suggests the heterogeneous set of ASD candidates share transcriptional networks related to synapse formation and elimination, protein turnover, and mitochondrial function.

Show MeSH

Related in: MedlinePlus

Transcriptome-wide molecular interaction network. a A heatmap of the expression of 13,563 genes (rows) across all 480 samples (columns). Samples are ordered first by brain region (color-code at the top) and then by age. The dendrogram to the right shows the clustering of all the genes into 32 modules. Modules with significant enrichment (p < 10−3) of genes from the ASD list are colored while other modules are shown in gray. b Enrichment of ASD candidate genes in each of the modules showing high significance in the magenta, brown, orange, and purple modules (represented by −log10(p), FDR-corrected)
© Copyright Policy - OpenAccess
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4644211&req=5

Fig3: Transcriptome-wide molecular interaction network. a A heatmap of the expression of 13,563 genes (rows) across all 480 samples (columns). Samples are ordered first by brain region (color-code at the top) and then by age. The dendrogram to the right shows the clustering of all the genes into 32 modules. Modules with significant enrichment (p < 10−3) of genes from the ASD list are colored while other modules are shown in gray. b Enrichment of ASD candidate genes in each of the modules showing high significance in the magenta, brown, orange, and purple modules (represented by −log10(p), FDR-corrected)

Mentions: The transcriptome-wide co-expression network was constructed from all genes expressed in the brain (13,563 genes), based on their expression profile across all samples (480 samples, i.e., all brain structures and developmental stages). Genes were hierarchically clustered based on Spearman’s rank correlation and complete linkage between pairs of genes. The resulting network consisted of 32 modules of varying size (from 36 to 1386 genes), as shown in Fig. 3a. Visual analysis of the heatmap and average expression patterns of member genes from each of the 32 modules demonstrated that none were specific to particular anatomical regions. This observation is consistent with the results from a similar dataset of human brain development assessed by microarray (Ben-David and Shifman 2012a). We did not observe any pre/postnatal-specific expression patterns in any of the 32 modules (Fig. S3). The genes comprising each of the 32 modules are listed in Table S5.Fig. 3


Shared Pathways Among Autism Candidate Genes Determined by Co-expression Network Analysis of the Developing Human Brain Transcriptome.

Mahfouz A, Ziats MN, Rennert OM, Lelieveldt BP, Reinders MJ - J. Mol. Neurosci. (2015)

Transcriptome-wide molecular interaction network. a A heatmap of the expression of 13,563 genes (rows) across all 480 samples (columns). Samples are ordered first by brain region (color-code at the top) and then by age. The dendrogram to the right shows the clustering of all the genes into 32 modules. Modules with significant enrichment (p < 10−3) of genes from the ASD list are colored while other modules are shown in gray. b Enrichment of ASD candidate genes in each of the modules showing high significance in the magenta, brown, orange, and purple modules (represented by −log10(p), FDR-corrected)
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4644211&req=5

Fig3: Transcriptome-wide molecular interaction network. a A heatmap of the expression of 13,563 genes (rows) across all 480 samples (columns). Samples are ordered first by brain region (color-code at the top) and then by age. The dendrogram to the right shows the clustering of all the genes into 32 modules. Modules with significant enrichment (p < 10−3) of genes from the ASD list are colored while other modules are shown in gray. b Enrichment of ASD candidate genes in each of the modules showing high significance in the magenta, brown, orange, and purple modules (represented by −log10(p), FDR-corrected)
Mentions: The transcriptome-wide co-expression network was constructed from all genes expressed in the brain (13,563 genes), based on their expression profile across all samples (480 samples, i.e., all brain structures and developmental stages). Genes were hierarchically clustered based on Spearman’s rank correlation and complete linkage between pairs of genes. The resulting network consisted of 32 modules of varying size (from 36 to 1386 genes), as shown in Fig. 3a. Visual analysis of the heatmap and average expression patterns of member genes from each of the 32 modules demonstrated that none were specific to particular anatomical regions. This observation is consistent with the results from a similar dataset of human brain development assessed by microarray (Ben-David and Shifman 2012a). We did not observe any pre/postnatal-specific expression patterns in any of the 32 modules (Fig. S3). The genes comprising each of the 32 modules are listed in Table S5.Fig. 3

Bottom Line: Autism spectrum disorder (ASD) is a neurodevelopmental syndrome known to have a significant but complex genetic etiology.Furthermore, we also constructed co-expression networks from the entire transcriptome and found that ASD candidate genes were enriched in modules related to mitochondrial function, protein translation, and ubiquitination.Overall, our multi-dimensional co-expression analysis of ASD candidate genes in the normal developing human brain suggests the heterogeneous set of ASD candidates share transcriptional networks related to synapse formation and elimination, protein turnover, and mitochondrial function.

View Article: PubMed Central - PubMed

Affiliation: Delft Bioinformatics Lab, Delft University of Technology, Delft, The Netherlands. a.mahfouz@tudelft.nl.

ABSTRACT
Autism spectrum disorder (ASD) is a neurodevelopmental syndrome known to have a significant but complex genetic etiology. Hundreds of diverse genes have been implicated in ASD; yet understanding how many genes, each with disparate function, can all be linked to a single clinical phenotype remains unclear. We hypothesized that understanding functional relationships between autism candidate genes during normal human brain development may provide convergent mechanistic insight into the genetic heterogeneity of ASD. We analyzed the co-expression relationships of 455 genes previously implicated in autism using the BrainSpan human transcriptome database, across 16 anatomical brain regions spanning prenatal life through adulthood. We discovered modules of ASD candidate genes with biologically relevant temporal co-expression dynamics, which were enriched for functional ontologies related to synaptogenesis, apoptosis, and GABA-ergic neurons. Furthermore, we also constructed co-expression networks from the entire transcriptome and found that ASD candidate genes were enriched in modules related to mitochondrial function, protein translation, and ubiquitination. Hub genes central to these ASD-enriched modules were further identified, and their functions supported these ontological findings. Overall, our multi-dimensional co-expression analysis of ASD candidate genes in the normal developing human brain suggests the heterogeneous set of ASD candidates share transcriptional networks related to synapse formation and elimination, protein turnover, and mitochondrial function.

Show MeSH
Related in: MedlinePlus