Unexpected links reflect the noise in networks

Abstract

BackgroundGene covariation networks are commonly used to study biological processes. The inference of gene covariation networks from observational data can be challenging, especially considering the large number of players involved and the small number of biological replicates available for analysis.ResultsWe propose a new statistical method for estimating the number of erroneous edges in reconstructed networks that strongly enhances commonly used inference approaches. This method is based on a special relationship between sign of correlation (positive/negative) and directionality (up/down) of gene regulation, and allows for the identification and removal of approximately half of all erroneous edges. Using the mathematical model of Bayesian networks and positive correlation inequalities we establish a mathematical foundation for our method. Analyzing existing biological datasets, we find a strong correlation between the results of our method and false discovery rate (FDR). Furthermore, simulation analysis demonstrates that our method provides a more accurate estimate of network error than FDR.ConclusionsThus, our study provides a new robust approach for improving reconstruction of covariation networks.ReviewersThis article was reviewed by Eugene Koonin, Sergei Maslov, Daniel Yasumasa Takahashi.Electronic supplementary materialThe online version of this article (doi:10.1186/s13062-016-0155-0) contains supplementary material, which is available to authorized users.

CVID enteropathy is characterized by exceeding low mucosal IgA levels and interferon-driven inflammation possibly related to the presence of a pathobiont

“…Second, we reasoned that since IgA is known to be protective against pathogens and can regulate commensal microbiota [16–18], bacterial species potentially contributing to disease in E-CVID patients would have increased abundance in E-CVID tissue compared to control tissue and would be negatively correlated with IgA levels. To select such bacterial candidates, we combined the results of correlations between bacterial abundances and IgA gene expression with the results of enrichment of bacteria in E-CVID using a recently developed statistical approach [19, 20]. Thus, out of 582 OTUs detected in duodenal biopsies, we found 45 OTUs (FDR<20%) that were enriched with bacteria potentially involved in enteropathy (red dots in right lower corners in both panels of Figure 3d, Supplementary Table S6).…”

“…The sign of correlation coefficients in three groups should be consistent (all positive correlation or all negative correlation across three groups). Also, we selected correlations that had sign of correlation consistent with possibility of causal relations between genes[19]. Third, human gene-gene network was generated by requiring all p-values of correlation coefficient from three groups <0.05 and combined FDR<0.1.…”

Transkingdom Networks: A Systems Biology Approach to Identify Causal Members of Host–Microbiota Interactions

“…Next, we identify the proportion of unexpected correlations (PUC) [50] (GD: line 83). If two elements have a regulatory relationship we expect them to behave in certain ways.…”

Differentially correlated genes in co-expression networks control phenotype transitions

“…Next, the pairs with false discovery rate (fdr) 55 lower than a threshold are chosen. At last, only the pairs that pass PUC 56 are considered correlated and therefore represent edges in the network.…”