Scientific research, so far, has been driven by star scientists working within epistemic communities. Transdisciplinarity (TD) has recently emerged as yet another vehicle for promoting scientific breakthroughs. TD science is based on the premise that scientific breakthroughs emerge through collaborative processes that integrate knowledge across disciplinary boundaries. However, this process is fraught with creative tensions as scientists from different "thought worlds" contribute in an asynchronous fashion to an ongoing initiative to obtain occasional breakthroughs. How such creative tension is managed can critically impact the outcomes of TD research. Despite the importance of TD science, there is only limited research on the processes that constitute such efforts. This study seeks to develop a new explanatory model of the architecture of TD research in which both its macro-structural and the micro-processual elements that enable translation across these disciplinary boundaries are examined. For instance, recent literature underscores the important roles that brokers play in bridging epistemic communities by facilitating required translation processes. Yet, little is known about how exactly brokers operate in these contexts. For instance, how do they resolve emerging conflicts and what roles do they play in governing TD research? How do brokers impact an accumulative process that requires long periods of sustained contributions from scientists who work in a distributed and parallel fashion, yet whose activities may be punctuated by occasional moments when breakthroughs occur? The first component of this study attempts to answer these questions. The inquiry frame that is used for this "science of science" project is appreciative of the accumulative yet fundamentally transformative nature of TD research. The research perspective recognizes that the underlying cumulative process is such that any progress that is made over time becomes inscribed into material artifacts, tools and techniques. The second part of the study, therefore, examines these material artifacts that serve as intermediate outcomes marking the progression of a complex and non-linear process. Equally importantly, these outcomes, as they accumulate, serve as platforms of resources shaping the activities of scientists who function in a distributed yet parallel fashion. The study investigates how this platform of material artifacts emerges as part of the research process and how it is transformed in use. The third part of the study attempts to understand the dynamic interactions between the macro structure and the micro-processes of collaboration through brokerage on the one hand and through translation on the other. Successful transdisciplinary research will, by definition, change the macro structure of collaboration. As this happens, the locus of brokerage will shift from one part of a network to another as critical problems are addressed and as other problems arise. Such a dynamic perspective on the architecture of collaboration has not yet been undertaken. This research presents a unique opportunity to study TD research collaboration in two settings that have the potential to yield valuable insights. In studying the National Institutes of Health's TTURC initiative, there is an opportunity to see how scientists from 7 different institutions collaborate over time to address emergent issues in a dynamic fashion to generate breakthroughs. The ATLAS/CERN provides the appropriate venue to study how 2,000 scientists and engineers at 151 institutions in 34 countries engage with one another to handle such creative tensions. In this way, the research directly advances one of SciSIP's objectives to understand the creativity process. Drawing on social network analyses and narrative approaches to understanding drivers of key events in collaborative endeavors, the study develops a new explanatory model of the process through which TD science unfolds over time.