My group continued to work on computational methods to study the dynamics of biological networks, impact of genetic and structural variation on gene expression and phenotype with emphasis on studies related to cancer and its heterogeneity. We also continued our research on the role of DNA conformational dynamics for gene regulation, and methods for analysis of HT-SELEX data. In particular, we continued to work on developing new computational methods to delineate genetic underpinnings of cancer and interactions between them. Utilizing the strength of our group in designing new algorithmic methods, our primary interest was in developing new network based based approaches (1). Knowledge about mutual exclusivity relationships can provide important insights into cancer drivers, cancer-driving pathways, and cancer subtypes (2). It can also be used to predict new functional interactions between cancer driving genes and novel cancer drivers. Until recently, most of mutual exclusivity analyses were performed focusing on a limited set of genes in part due to the computational cost required to rigorously compute all pairwise p-values. To address this limitation, we developed Weighted Sampling Mutual Exclusivity (WeSME) a new efficient method to estimate p-values while controlling the mutation rates of individual patients and genes similar to the permutation test (3). Given the efficiency of our proposed method, we expect that it will be widely used. We also continue to develop our comprehensive software package, AptaTools, for analysis of HT-SELEX derived Aptamers. Aptamers are synthetic but biologically active single-stranded (ribo)nucleic molecules, typically ranging between 15 and 120 nucleotides. These short sequences can be designed to bind, with high affinity and specificity, a vast spectrum of molecular targets spanning from small organic molecules over macromolecules such as proteins to whole viruses. Aptamers and apatamer based conjugates are increasingly considered as potential RNA based drugs. My group has previously developed a number of computational tools to tackle the challenges of Aptamer research. We have now significantly expanded our toolbox (4,5,6). Our most important contribution in the reporting priod was AptaTRACE a method for detecting sequence-structure motifs by identifying sequence motifs undergoing selection during HT-SELEX procedure towards a particular secondary structure context. With the help from our experimental collaborators we showed that sequence motifs identified by AptaTRACE are important for aptamer binding (6). AptaTRACE has the potential to become an important tool for aptamer design and thus the publication of this method in Cell Systems was covered by NIH press release. Continuing our long-standing collaboration with Brian Oliver's group on gene regulation in Drosophila. We profiled expression of the genes with gene dose was reduced from two to one in the engineered flies. While expression of most one-dose genes was reduced, the gene-specific dose responses were heterogeneous. In both genetic backgrounds that we analysed, we observed a clear reduction in gene expression from one as compared to two copies. Our analysis confirmed previous reports that reduced expression is not 2-fold. We observed a mean 1.1-fold compensation against gene dose reduction. As in previous work, we observed that compensation was not due to a uniform effect on all genes, as one copy gene expression was skewed towards compensation. These data indicate that different genes show differences in compensation responses, continuously ranging from common modest compensation levels through to more rare nearly perfect compensation, with infrequent extreme deviations (7). We are also continuing our successful collaboration with David Levens focusing on the role of DNA conformational dynamics in gene regulation (8,9). In particular, to elucidate the role of Pol II poising in B cell activation, we recenlty compared Pol II profiles in resting and activated B cells. We found that while Pol II poised genes generally overlap functionally among different B cell states and correspond to the functional groups previously identified for other cell types, non-poised genes are B cell state specific. Focusing on the changes in transcription activity upon B cell activation, we found that the majority of such changes were from poised to non-poised state. The genes showing this type of transition were functionally enriched in translation, RNA processing and mRNA metabolic process. Interestingly, we also observed a transition from non-poised to poised state. Within this set of genes we identified several Immediate Early Genes (IEG), which were highly expressed in resting B cell and shifted from non-poised to poised state after B cell activation. Thus Pol II poising does not only mark genes for rapid expression in the future, but it is also associated with genes that are silenced after a burst of their expression. Finally, we performed comparative analysis of the presence of G4 motifs in the context of poised versus non-poised but active genes. Interestingly we observed a differential enrichment of these motifs upstream versus downstream of TSS depending on poising status. The enrichment of G4 sequence motifs upstream of TSS of non-poised active genes suggests a potential role of quadruplexes in expression regulation. In addition to these topics central to our main line of reserach, we continue to collaborate with other groups inside and outside NIH by contributing our computational biology expertise to their studies (10, 11,12, 13) and support computational biology community by chairing two key Computational Biology conferences (14,15).
Showing the most recent 10 out of 72 publications
