Adaptively inferring human transcriptional subnetworks

There is a News and Views associated with this document.

Debopriya Das^1,a, Zaher Nahlé² & Michael Q Zhang¹

^aPresent address: Life Sciences Division, Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA

Article highlights

Synopsis

The importance of achieving an accurate quantitative understanding of gene regulation in humans can hardly be overstated. Deregulation of gene expression is a recurring theme in development and progression of several diseases including cancer. The emergence of new experimental platforms that probe transcription globally promises a comprehensive view of these fundamental biological processes in a large number of mammalian systems, in which very little is known in terms of their transcriptional regulation. By integrating the expression profiles with the genomic sequence information computationally, it is now possible to obtain a snapshot of the active transcriptional subnetworks in lower eukaryotes with a reasonable accuracy (Das et al, 2004; Wang et al, 2005). We define a transcriptional subnetwork as the set of transcription factors (TFs) as represented by the combinations of cognate cis-regulatory motifs, their target genes and the physiological processes they regulate (Figure 2). The generalization of such approaches to mammals remains challenging however (see, e.g., Figure 1 in Tompa et al, 2005). This is due to multiple factors, including enhanced degeneracy of TF binding sites, significantly elevated role of interactions between TFs in promoter recognition and multicellular architecture of mammals. Current computational methods, which are primarily clustering-based, do not adequately address these complicating factors. Moreover, many genes do not cluster tightly enough that their regulatory motifs can be discovered reliably. There is also marked subjectivity in how targets are determined. This work presents a minimally biased approach motivated by the switch-like behavior of transcriptional response, which overcomes the aforementioned limitations. It identifies potentially active motif combinations in proximal promoters by examining their correlation with mRNA expression levels across the genes. In this approach, both the active motif combinations and their target genes are learnt directly from the expression data in a condition-specific manner, and thus, adaptively. We demonstrate that this method can systematically infer transcriptional subnetworks in mammals from expression data with accuracy similar to those obtained for lower eukaryotes.

Figure 2

Schematic representation of our analysis. A snapshot of the tissue-specific transcriptional subnetworks discovered from microarray data on adult human liver under a normal condition.

Full figure and legend (156K)Figures & Tables index

We applied our algorithm to the expression profile of adult human liver measured under a normal condition (Su et al, 2004) and discovered three functional liver-specific motif combinations. The inferred model was used to obtain their target genes, a Gene Ontology enrichment analysis of which subsequently revealed the over-represented biological pathways, thus leading to transcriptional subnetworks active in the profiled sample (Figure 2). HNF-1, the pleiotropic regulator of liver-specific genes, is among the three liver-specific combinations. We observe, a posteriori, that >70% of our predicted HNF-1 targets have been previously validated in biochemical assays. The other two liver-specific combinations are novel, one of which regulates sugar metabolism pathways, and another regulates lipid transport and metabolism. There are certain other advantages to this approach. For instance, we are able to identify the mRNA mixing effects present in tissue samples derived from a whole organ such as the liver. Additionally, we notice that several targets achieve their maximum expression in a tissue different from where the motif combination has its maximal regulatory effect. This suggests that genes are coregulated across multiple tissues, as one would expect in a synexpression group (Niehrs and Pollet, 1999). A distinct advantage of our method is that expression profiles from only a few conditions are necessary to reach this conclusion.

TFs regulate genes in a condition-specific manner. Hence, a particular TF can activate different sets of genes under different conditions (Zhu et al, 2005a). Application to human cell-cycle data (Whitfield et al, 2002) revealed that this technique can model such condition-specific gene regulation. Namely, the predicted targets of E2F in G1/S and G2/M phases are significantly different, as one would expect biologically (Zhu et al, 2005a). This is a natural outcome of the fact that the TF binding thresholds are learnt directly from response data in our approach. Many identified targets have been previously characterized as direct E2F targets, and the novel targets display strong E2F binding characteristics. We experimentally examined two G2/M-specific novel targets involved in hepatocellular carcinomas, CDC16 and DLG7, using an inducible E2F system. Our experiments confirmed that DLG7 and CDC16 are indeed direct E2F targets, in accordance with our computational predictions. The role of E2F as a G2/M activator has only been recently demonstrated. Our findings reaffirm and extend this hypothesis. Furthermore a strong correlation between the mitotic control and tumor progression in liver has been previously noted (Tsou et al, 2003). Intriguingly, our study suggests that part of this control may be exerted by E2F.

One of the major challenges in modern biology is to comprehensively decipher the regulatory network architecture within humans, both for advancement of fundamental understanding and in the development of novel therapeutics. We have presented an unsupervised algorithm to achieve this goal. It directly accounts for degeneracy and interactions among cis-regulatory elements, the two key impediments to modeling mammalian transcription, and leads to concrete testable hypotheses for transcriptional end points of pathways that are active under any specific biological condition. We find that, on average, the predicted models can explain 24% of the variation in expression data for tissue-restricted profiles and 21% for temporal profiles, comparable to those achieved in lower eukaryotes (Das et al, 2004). In addition, as our study clearly demonstrates, correlation with expression is a promising way to identify direct TF targets. By contrast, it is quite challenging to discriminate direct targets from indirect ones using previous approaches (Kirmizis and Farnham, 2004). Furthermore, our method for target determination does not depend on arbitrary fold cutoffs invoked by many comparable methods. Consequently, we can detect bona fide targets that undergo very subtle changes in expression as well. Autoregulatory loops, prevalent in biological networks, can also be easily discovered. The algorithm is equally applicable to both expression and ChIP-chip data. In summary, we think this approach will help accelerate systematic understanding of how phenotypic complexity is regulated at the transcription level in a wide range of mammalian systems and make scope for novel therapeutic interventions.

MORE ARTICLES LIKE THIS

These links to content published by NPG are automatically generated.