Limits...
Multiple Transcriptome Data Analysis Reveals Biologically Relevant Atopic Dermatitis Signature Genes and Pathways.

Ghosh D, Ding L, Sivaprasad U, Geh E, Biagini Myers J, Bernstein JA, Khurana Hershey GK, Mersha TB - PLoS ONE (2015)

Bottom Line: Keratinocytes are known to play a major role in barrier function due to their location in the epidermis.Our result suggests that besides immune- mediated pathway, skin barrier pathways such as the keratinocyte differentiation pathway play a key role in AD pathogenesis.A better understanding of the role of keratinocytes in AD will be important for developing novel "barrier therapy" for this disease.

View Article: PubMed Central - PubMed

Affiliation: Division of Immunology, Allergy & Rheumatology, Department of Internal Medicine, University of Cincinnati, Cincinnati, United States of America.

ABSTRACT
Several studies have identified genes that are differentially expressed in atopic dermatitis (AD) compared to normal skin. However, there is also considerable variation in the list of differentially expressed genes (DEGs) reported by different groups and the exact cause of AD is still not fully understood. Using a rank-based approach, we analyzed gene expression data from five different microarray studies, comprising a total of 127 samples and more than 250,000 transcripts. A total of 89 AD gene expression signatures '89ADGES', including FLG gene, were identified to show dysregulation consistently across these studies. Using a Support Vector Machine, we showed that the '89ADGES' discriminates AD from normal skin with 98% predictive accuracy. Functional annotation of these genes implicated their roles in immune responses (e.g., betadefensin, microseminoprotein), keratinocyte differentiation/epidermal development (e.g., FLG, CORIN, AQP, LOR, KRT16), inflammation (e.g., IL37, IL27RA, CCL18) and lipid metabolism (e.g., AKR1B10, FAD7, FAR2). Subsequently, we validated a subset of signature genes using quantitative PCR in a mouse model. Using a bioinformatic approach, we identified keratinocyte pathway over-represented (P = <0.0006) among the 89 signature genes. Keratinocytes are known to play a major role in barrier function due to their location in the epidermis. Our result suggests that besides immune- mediated pathway, skin barrier pathways such as the keratinocyte differentiation pathway play a key role in AD pathogenesis. A better understanding of the role of keratinocytes in AD will be important for developing novel "barrier therapy" for this disease.

Show MeSH

Related in: MedlinePlus

Distribution of fold change and average percent significance rank of 89ADEGS.Identification and ranking of differentially expressed genes (DEGs) from individual studies using both ranked statistical significance (x-axis) and biological relevance (fold change, FC, y-axis) provide better understanding of relevant gene sets than either of the methods. The p-values of the 89ADGES member genes were arranged (smaller to larger p-values) in each of the datasets and percent-ranked. For example, the percent-rank of a DEG/transcript at the 500th position within a set of 50,000 transcript will rank 5th. Average percent ranks each 89ADGES member gene was determined by averaging the percent-ranks from 5 datasets and plotted against respective average fold-change values. A lower percent rank indicates higher significance (Please see S2 Table for details). Upregulated genes are in red color.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4696650&req=5

pone.0144316.g003: Distribution of fold change and average percent significance rank of 89ADEGS.Identification and ranking of differentially expressed genes (DEGs) from individual studies using both ranked statistical significance (x-axis) and biological relevance (fold change, FC, y-axis) provide better understanding of relevant gene sets than either of the methods. The p-values of the 89ADGES member genes were arranged (smaller to larger p-values) in each of the datasets and percent-ranked. For example, the percent-rank of a DEG/transcript at the 500th position within a set of 50,000 transcript will rank 5th. Average percent ranks each 89ADGES member gene was determined by averaging the percent-ranks from 5 datasets and plotted against respective average fold-change values. A lower percent rank indicates higher significance (Please see S2 Table for details). Upregulated genes are in red color.

Mentions: Fig 2 shows unsupervised hierarchical cluster analysis (HCA) (panel A) and Principal component analysis (Panel B) of the datasets. In HCA, euclidean distance and complete linkage were used to obtain similarities/dissimilarities among sets of individuals according to their expression values of all 89 genes. Branch height represents dissimilarity. As expected, samples were clustered into two groups: AD patients versus healthy individuals. Principal component analysis (PCA) of the entire AD and control samples are shown in Fig 2 (Panel B). The first PC accounted for more than double the variance of the second PC. Further analysis using principal component scatter plot showed that the first two PCs contributed to 60–80% of the total variation in AD. Fig 3 shows log average fold-changes of expression of 89DEGs plotted against p-value ranks. Result shows that the top significant genes belong to keratinocyte differentiation, cell morphology as well as some innate immune genes, which were consistently down-regulated.


Multiple Transcriptome Data Analysis Reveals Biologically Relevant Atopic Dermatitis Signature Genes and Pathways.

Ghosh D, Ding L, Sivaprasad U, Geh E, Biagini Myers J, Bernstein JA, Khurana Hershey GK, Mersha TB - PLoS ONE (2015)

Distribution of fold change and average percent significance rank of 89ADEGS.Identification and ranking of differentially expressed genes (DEGs) from individual studies using both ranked statistical significance (x-axis) and biological relevance (fold change, FC, y-axis) provide better understanding of relevant gene sets than either of the methods. The p-values of the 89ADGES member genes were arranged (smaller to larger p-values) in each of the datasets and percent-ranked. For example, the percent-rank of a DEG/transcript at the 500th position within a set of 50,000 transcript will rank 5th. Average percent ranks each 89ADGES member gene was determined by averaging the percent-ranks from 5 datasets and plotted against respective average fold-change values. A lower percent rank indicates higher significance (Please see S2 Table for details). Upregulated genes are in red color.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4696650&req=5

pone.0144316.g003: Distribution of fold change and average percent significance rank of 89ADEGS.Identification and ranking of differentially expressed genes (DEGs) from individual studies using both ranked statistical significance (x-axis) and biological relevance (fold change, FC, y-axis) provide better understanding of relevant gene sets than either of the methods. The p-values of the 89ADGES member genes were arranged (smaller to larger p-values) in each of the datasets and percent-ranked. For example, the percent-rank of a DEG/transcript at the 500th position within a set of 50,000 transcript will rank 5th. Average percent ranks each 89ADGES member gene was determined by averaging the percent-ranks from 5 datasets and plotted against respective average fold-change values. A lower percent rank indicates higher significance (Please see S2 Table for details). Upregulated genes are in red color.
Mentions: Fig 2 shows unsupervised hierarchical cluster analysis (HCA) (panel A) and Principal component analysis (Panel B) of the datasets. In HCA, euclidean distance and complete linkage were used to obtain similarities/dissimilarities among sets of individuals according to their expression values of all 89 genes. Branch height represents dissimilarity. As expected, samples were clustered into two groups: AD patients versus healthy individuals. Principal component analysis (PCA) of the entire AD and control samples are shown in Fig 2 (Panel B). The first PC accounted for more than double the variance of the second PC. Further analysis using principal component scatter plot showed that the first two PCs contributed to 60–80% of the total variation in AD. Fig 3 shows log average fold-changes of expression of 89DEGs plotted against p-value ranks. Result shows that the top significant genes belong to keratinocyte differentiation, cell morphology as well as some innate immune genes, which were consistently down-regulated.

Bottom Line: Keratinocytes are known to play a major role in barrier function due to their location in the epidermis.Our result suggests that besides immune- mediated pathway, skin barrier pathways such as the keratinocyte differentiation pathway play a key role in AD pathogenesis.A better understanding of the role of keratinocytes in AD will be important for developing novel "barrier therapy" for this disease.

View Article: PubMed Central - PubMed

Affiliation: Division of Immunology, Allergy & Rheumatology, Department of Internal Medicine, University of Cincinnati, Cincinnati, United States of America.

ABSTRACT
Several studies have identified genes that are differentially expressed in atopic dermatitis (AD) compared to normal skin. However, there is also considerable variation in the list of differentially expressed genes (DEGs) reported by different groups and the exact cause of AD is still not fully understood. Using a rank-based approach, we analyzed gene expression data from five different microarray studies, comprising a total of 127 samples and more than 250,000 transcripts. A total of 89 AD gene expression signatures '89ADGES', including FLG gene, were identified to show dysregulation consistently across these studies. Using a Support Vector Machine, we showed that the '89ADGES' discriminates AD from normal skin with 98% predictive accuracy. Functional annotation of these genes implicated their roles in immune responses (e.g., betadefensin, microseminoprotein), keratinocyte differentiation/epidermal development (e.g., FLG, CORIN, AQP, LOR, KRT16), inflammation (e.g., IL37, IL27RA, CCL18) and lipid metabolism (e.g., AKR1B10, FAD7, FAR2). Subsequently, we validated a subset of signature genes using quantitative PCR in a mouse model. Using a bioinformatic approach, we identified keratinocyte pathway over-represented (P = <0.0006) among the 89 signature genes. Keratinocytes are known to play a major role in barrier function due to their location in the epidermis. Our result suggests that besides immune- mediated pathway, skin barrier pathways such as the keratinocyte differentiation pathway play a key role in AD pathogenesis. A better understanding of the role of keratinocytes in AD will be important for developing novel "barrier therapy" for this disease.

Show MeSH
Related in: MedlinePlus