Sridhar Hannenhalli - US grants

[unreadable] DESCRIPTION (provided by applicant): A broad goal in genome research is to better understand transcriptional regulation. We propose to complement current experimental efforts towards this goal through computational approaches. Gene transcription is regulated by a network of transcription factors (TF); to accomplish their task, TFs bind to specific DNA elements in the relative vicinity of the gene, interact with each other and with polymerase. The overall task of transcriptional control is divided among smaller groups of closely functioning TFs, or transcriptional modules, such that each module regulates transcription in response to specific stimuli or environmental condition. This division of control provides a modular mechanism to co-regulate groups of functionally related genes. The combinatorial interactions among TFs and the DNA elements that facilitate these interactions motivate the algorithmic approaches to study transcriptional regulation. We will (1) develop an EM approach to simultaneously detect transcription factor binding motifs and their interacting partners from genome-wide ChIP experiments; we will apply this to genome-wide yeast ChlP-chip data and to genome-wide CREB binding data in rat, (2) develop novel graph-theoretic approaches to detect transcriptional modules, and apply the methods to detect modules driving transcription in two biological processes - long term memory storage and heart failure - in collaboration with experimentalists, (3) develop novel Gibbs sampling approach to detect dense sub-graphs in a multi-partite graph as a means to identify modules, and apply this to detect modules driving tissue-specificity in human. [unreadable] [unreadable] [unreadable]

DESCRIPTION (provided by applicant: FOX family of transcription factor genes have expanded to over 40 members in mammals, and are involved in a variety of key developmental processes and human diseases. However, several key question concerning their evolutionary expansion, functional elaboration, transcriptional regulation, and selection are not completely understood. Gene duplication is a major driver of evolutionary innovation, as it allows an organism to elaborate existing biological function with specialization or diversification, while at the same time potentially avoiding negative fitness effects. Elaboration of new or specialized gene functions can involve at least two distinct pathways: (1) alteration of the spatial and temporal expression pattern of the gene, (2) alteration of the interaction partners, both DNA and protein. Alteration of expression patterns is likely to involve evolutionary divergence of the upstream cis regulatory elements while the alteration of interaction partners may involve sequence changes in functional domains of the protein. Previous comparative and evolutionary studies have focused on conserved aspects of gene sequence, structure and networks. Part of the challenge in studying divergence is that it is difficult to distinguish functionally relevant divergence from neutral drift. We propose to develop computational algorithms and statistical analysis methods that key in on correlated patterns of evolutionary divergence to identify novel regulatory elements and help elucidate the evolution of gene family members and their relation to disease phenotypes. In specific aim 1 we will develop a novel method - Intra-genomic Phylogenetic Footprinting - to identify cis elements by exploiting correlated patterns of divergence in the regulatory-sequence and the expression of paralogs. In specific aim 2, we will investigate additional pathways by which gene paralogs diversify in function and characterize the relationships between these pathways. In specific aim 3, we will develop methods to quantify selective pressure and epistatic interactions among human cis regulatory polymorphisms. The methods will be comprehensively applied to gene families in multiple species. Application to specific gene families will provide insights into their functional expansion. Selected predictions will be experimentally validated. Besides providing a broader evolutionary understanding of gene family expansion, transcriptional regulation, and intraspecific variability in human populations, specific applications of the proposed research will be of direct biomedical relevance. PUBLIC HEALTH RELEVANCE: FOX family of transcription factor genes have expanded to over 40 members in mammals, and are involved in a variety of key developmental processes and human diseases. However, several key question concerning their evolutionary expansion, functional elaboration, transcriptional regulation, and selection are not completely understood. We proposed to develop computational algorithms and statistical analysis methods to address some of these questions concerning the evolution of gene families and transcriptional regulation.

DESCRIPTION (provided by applicant): Heart failure has become the most common reason for adult hospitalization in the industrialized world. Basic research has shown that pathologic stresses promote heart failure via activation of cardiac transcription factors (TFs). These TFs interact with each other and with co-activators to form "transcriptional modules" (TMs) that alter cardiac gene expression to cause myocyte hypertrophy and failure. In light of their critical role in heart failure pathogenesis, TMs have been the subject of intensive research in animal models over the past decade. By contrast, the role of TMs in human heart failure remains largely unexplored because of limited methods for studying TFs in human subjects. We have piloted computational approaches that enable assessment of TF function in the failing human heart by integrating cardiac gene expression data from microarray experiments with readily available genome sequence data. However, these do not sufficiently address the complexity of the underlying biology, and more rigorous methods are needed. The purpose of this exploratory/developmental research proposal is to determine transcriptional modules (TMs) associated with human heart failure using a refined computational approach and to experimentally validate the most promising TM using standard in vitro techniques. We will develop novel integrative approaches to determine TMs associated with human heart failure and apply them to the Penn Cohort, the largest study of human cardiac gene expression in the published literature. These approaches integrate whole genome expression data with data from promoter sequences, TF binding motifs, and cross-species sequence alignments. Our most promising newly identified heart failure TM will be validated experimentally using in vitro reporter gene assays in cardiac myocytes. This interdisciplinary proposal will build on existing collaboration between a clinical investigator, a computational biologist, a biostatistician, and a molecular biologist. Our research will determine TMs directly relevant to the pathogenesis of human heart failure. In doing so, we will extend an important body of work in animal models to the arena of clinical investigation. The specific heart failure TMs we identify will become the focus of future research performed by our own group and by others. These studies may ultimately lead to new drugs that target transcriptional mechanisms of hypertrophy, a central feature of heart failure that is not directly targeted by any current therapy. Lastly, the computational methods we develop should have broad application to study other human diseases. Heart failure has become the most common reason for adult hospitalization in the industrialized world. Research performed over the past decade has determined that cardiac transcription factors play a crucial role in the pathogenesis of heart failure, but these findings have not been extended to human subjects. This proposal will develop and apply novel genomic approaches to study the role of cardiac transcription factors in human subjects with advanced heart failure.

"This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5)."

The University of Maryland College Park is awarded a grant to develop new algorithms and a suite of software tools based on a general and flexible definition of a "network module" in order to extract meaningful biological clusters from noisy and incomplete protein-protein interaction data. Recently developed high-throughput techniques are being used to sample protein-protein interactions from many organisms and are creating a wealth of data that must be analyzed computationally. A central challenge in the study of these networks is finding biologically meaningful and interpretable modules within them. The new tools and algorithms will be used to improve visualization of protein interaction networks, identify protein complexes and biological processes embedded within the network data, and to discover redundant pathways from synthetic lethal interaction data. They will also be applied to comparing the interaction networks of several different species. The resulting network analysis software will expand the capabilities of both systems biologists and biologists working on particular protein complexes and pathways to make better use of noisy network data, and the proposed visualization software will vastly improve researchers? capability to interactively explore these networks. A public database will be created to curate computationally predicted annotations made by this, and other, projects using this method. The work will strengthen understanding of the organization of biological networks, and it will have broader impact by increasing the information technology infrastructure available for the analysis of interaction data, providing better transfer of hypotheses between computational biologists and biologists, and by the training of undergraduates in a summer internship program. Information about the project and how to access the database and software will be available at the project website http://www.cbcb.umd.edu/research/bionet/.

DESCRIPTION (provided by applicant): A mechanistic understanding of eukaryotic gene transcription is an important long-term goal in biology as it pertains to human health. Biotechnological advances have revealed heretofore unknown complexity of transcriptional regulation, challenging current models and raising new questions. The proposed projects address three such questions via novel methods and analysis, and promise to enhance our understanding of transcriptional control. (1) There are now a several known examples of network rewiring where a group of genes have conserved expression over long evolutionary distances but the transcriptional mechanisms underlying the expression of the genes has diverged. Understanding the mechanisms for such rewiring has implications for our understanding of evolvability and robustness of organisms. In the specific aim 1, we will develop computational methods to identify instances of transcriptional network rewiring and characterize the conditions facilitating the rewiring. (2) While traditionally, a particular transcription facto (TF) was believed to bind to a specific DNA motif, now it is becoming apparent that many TFs may recognize distinct motifs that modulate functionally distinct outcomes. In the specific aim 2, we will develop computational methods to discover and characterize functional subclasses of transcription factor binding sites. (3) Many important developmental enhancers act from a distance, up to a million nucleotides away from the target gene. How the enhancers accomplish their action-at-a-distance is not entirely clear and has implications for our understanding of developmental and tissue-specific gene regulation. In the specific aim 3 we will develop methods to map enhancers to their distal target genes. 1

This proposal seeks funding for travel support for students to participate in the ACM BCB (Bioinformatics and Computational Biology) Conference in September 2013 in Washington D.C. The ACM-BCB is expected to be a premier forum in Bioinformatics and Computational Biology. The conference will host three keynote sessions, research sessions, session-specific invited speakers, workshops, tutorials, panel discussion, and poster sessions. Special activities such as a doctoral consortium and student poster session will be specifically targeted for the students. The ACM-BCB conference has been successful in building a bioinformatics and computational biology community through the newly established ACM Special Interest Group in Bioinformatics and Computational Biology (SIGBioinformatics). The proposed grant will broaden the participation graduate students and help develop the next generation of researchers and educators in the computer science, biological science, biomedical science and other related areas.

Co-PI: Sridhar Hannenhalli, University of Maryland
Co-PI: Chris Dardick, Appalachian Fruit Research Laboratory, USDA-ARS
Co-PI: Ann Callahan, Appalachian Fruit Research Laboratory, USDA-ARS

Fruits that are so enjoyed by people are also especially adapted for plant seed dispersal. Even closely related plants display different types of fleshy fruits that accomplish dispersal through variation in color, shape, size, texture and other traits. A major question for plant biologists and fruit crop scientists is how these diverse fruit types evolved and what factors control their development. If scientists could explain the molecular basis of different fruit types, new breeding tools could be used to increase the value of fruit crops and basic research could be advanced. The Rose family (Rosaceae) is an excellent example of how fruits have been selected and adapted to different types of dispersal. Members of the family, including the familiar strawberries, apples, and roses, all have unique fruits and floral structures from which the fruits are derived. This project investigates how these different fruits develop at the molecular level. Using comparative genomic and bioinformatics methods, the research will identify shared and unique genes that control the pattern of fruit development among five representatives of the Rose family: strawberries, peach, plum, apple and raspberry. Computational methods will be used to identify regulatory gene networks that explain how and when the diverse fruits form. High school teachers will be brought into the research program during the summer to study fruit science and computational data analysis, and curricula will be developed for use in the high school classroom. By integrating teachers into the research, the next generation of students will be impacted broadly and will gain hands-on exposure to genomics and bioinformatics.

The project investigates the molecular genetic mechanisms that underlie the phenotypic diversity of fleshy fruits in the Rosaceae. By leveraging the evolutionary conservation of genes within the Rosaceae genome, the power of comparative genomics will be used to identify key genes and mechanisms of fleshy fruit development. The recently published Rosaceae genome sequences (apple, strawberry, peach, raspberry and plum) provide an unprecedented opportunity to investigate this fundamental biological question. First, detail morphological and histological changes during floral and early stage fruit development will be characterized for the chosen Rosaceae species. This knowledge will guide sample collection for RNA-sequencing methods immediately before and after fertilization in specific floral tissues. Innovative bioinformatic tools and strategies will be developed to identify toolkit genes and regulatory gene networks, whose spatial or temporal shifts of expression may alter fleshy fruit programs in different species. Functional tests including RNAi, CRISPR, and tissue-specific ectopic expression will be carried out to test predicted toolkit genes using the established transformation systems in strawberry, FasTrack plum, and apple. The ultimate goal is to identify causal relationships between changes of toolkit gene expression and evolution of distinct fleshy fruit types. The research activities will be integrated with several educational and outreach programs including training of students and postdoctoral researchers in genomics and informatics, bioinformatics workshops for high school teachers from rural West Virginia and predominantly minority-serving schools of Prince Georges County of Maryland, and outreach to the public about Genetically Modified Organisms and fruit genomics. All sequence data will be made available through the public NCBI database and through the community database for the Rosaceae.

Proteins are one of the basic building blocks of living organisms. Each tissue carefully regulates the set of proteins that it produces, both of the types of protein made and the amounts of each. For instance, the proteins that make up the structure and enable the function of the heart are different from those defining the lungs or brain. A deep understanding of the processes that control what proteins are made and how this is timed is of fundamental scientific importance, with implications throughout biology. This project focuses on two critical points in protein regulation: the step in which genes are read off of the genomic DNA to make the intermediate mRNA (called transcription), and the step in which some pieces of the mRNA are removed and the ends reconnected (called splicing), so that translation to protein will result in versions with distinct properties. Complete profiling of mRNA produced by tissues is possible but results in very large, complex data sets. This project will create software tools to facilitate efficient analyses of these data sets, making sense of the two processes described above. The results can be applied to investigate a wide variety of biological questions. Besides the scientific contributions, several educational and outreach activities are aimed at students at the high school and undergraduate levels. In particular, a summer workshop will be organized to provide hands-on bioinformatics experiences to local high school teachers and representative students. Students in the University of Maryland Terrapin Teacher program will be mentored to create lesson plans in Bioinformatics; the lessons will subsequently be delivered at local high schools. In addition, high school students will be invited to participate in summer research experiences under the mentorship of the PIs.

Cellular morphology and function is determined by precise regulation of gene activity. A long-term goal of biological research is to fully decipher the mechanisms of gene regulation. Spatial and temporal regulation of final gene products is primarily executed when genes are transcribed and the resulting RNA is processed. In turn, transcriptional regulation is mediated by regulatory elements, such as enhancers and promoters, which are characterized by diffuse clusters of short degenerate DNA motifs. An important task in the analysis of the transcriptional regulation involves comparison of regulatory regions, both within and across genomes. With regards to post-transcriptional processing events, critical first steps are identification, characterization and quantification of alternatively processed variants of a gene. Increasingly, sequence data are available that hold the answers to these questions, but which must be mined in order to extract their secrets. The scale of the task of incorporating the massive amounts of next generation sequencing data render conventional solutions inefficient and thus presents a major computational bottleneck. The proposed research will develop efficient algorithms and tools for the analysis of gene expression at both transcriptional and post-transcriptional levels by exploiting recently developed ultrafast data structures for DNA words or k-mers. The proposed solutions to two related but different fundamental problems - quantification of regulatory region similarity, and quantification of alternative isoforms of a gene, provide an alternative to traditional approaches by reformulating the problems in terms of k-mer similarity searches, for which extremely efficient solutions have been recently developed and exploited for related problems, including transcript quantification (e.g. Sailfish). The proposed tools will enable investigation of certain fundamental mechanistic and evolutionary questions pertaining to transcriptional regulation and analysis of splice isoform variants at an unprecedented scale.
The research goals will be tied to several STEM educational activities focused on local high school students, both in terms of early research involvement and classroom education. Some of these activities will be organized at the Center for Bioinformatics and Computational Biology (CBCB), thereby encouraging the close interaction between the students and the center faculty.
The link to the results will be provided at PI's lab page at cbcb.umd.edu/~sridhar/software.html