Setting the Vision
Biomedicine has become an information science. In all areas of basic biomedical and clinical research, as well as in clinical practice, intelligent management of large, multidimensional data sets and the mining for meaningful patterns is becoming a key task. The data types in biomedicine are diverse. They range from qualitative recordings of patient behavior, to genome sequences, protein expression and activity profiles, stochastic time-series, to still and live images over many spatial and temporal scales. Yet, the methods used to manage, mine, and model these data build on the same statistical and mathematical foundations. In somewhat superficial terms, the informatics challenge in biomedicine is one of detecting, learning, and projecting patterns
Independent of the data type used, the central task in any biomedical study is to identify meaningful correlations between datasets that allow then the definition of the likelihood of association between observables, e.g., between a genomic sequence and disease risk, or between the spatial distribution of protein expression and cancer progression. The same association problem arises also with the building of predictive models. Here, correlations are sought between an experimental dataset as one observable, and the outcome of a model simulation as the other observable. Thus, innovation in biomedicine depends on innovation in detecting complex associations between increasingly higher-dimensional datasets.
Providing this innovation is the core task of bioinformatics. The task is largely independent of the field of application. In fact, the ultimate goal of bioinformatics must be to combine all data types into a global association scheme that links patient behavior and outcomes to disease risk and progression, as probed by the full spectrum of available measurements
Our definition of goals departs from the often narrower understanding of bioinformatics as the ‘computational branch of genetics and genomics’. While we are not the first to recognize that bioinformatics is becoming an overarching science of biomedical information processing, the launch of a new academic medical center department without ties to historical preconceptions of the term 'bioinformatics' provides us with an opportunity to establish a transformative intellectual and technical infrastructure that deploys to biomedicine innovative information science in the broadest sense.
The abundance of information science in virtually every clinical and basic research project generates an opportunity for the Department to become an integrator unit in the center of UTSW’s research mission.
The Department was established in Spring 2015 and interacts with the UTSW Campus along four interfaces:
- Core 'innovator' faculty and 'integrator' faculty with joint appointments. The Department is composed of scientists with research programs in foundational computer science topics and scientists who will translate this core innovation into the biomedical research arena by leading ‘hybrid’ experimental and computational research programs. Such ‘hybrid’ programs will require co-affiliation of computational scientists with clinical or basic science departments providing experimentalist "wet lab' infrastructure. This interdependence facilitates open conduits of communication across the campus.
- Bioinformatics Core Facility. The Department operates a core facility that serves the UTSW community with expertise in the use of established and cutting edge (home-built) computational solutions to biomedical data processing needs. Interactions with investigators across campus occur in a range of formats, from help desk consultations to educational programs (e.g., nanocourses), and formal grant funded collaborations (see business plan, BICF Overview). The core facility also offers a Fellows Program, in which graduate students and postdoctoral fellows from labs across campus reside in the BICF space for a limited amount of time to work on data processing problems.
- Collaborative research programs. To initiate collaborations on campus that turn into long-term project partnerships with extramural funding, the Department has set aside from its startup budget financial resources for 5 – 10 pilot grants linking a laboratory in the Department of Bioinformatics with a laboratory in any of other UTSW department. The primary criterion for funding will be demonstration of innovation in both informatics and experimental procedures.
- Graduate program in bioinformatics. The Department will take leadership in building a graduate program in bioinformatics. By deploying bioinformatics-trained students across campus, the graduate program will strive to raise the overall quality of data-rich research at UTSW. While this program is in development, the Department co-leads the graduate Computational and Systems Biology specialty track.
The long-term impact of bioinformatics at a medical school hinges on true innovations in core areas of computer science (CS) and applied mathematics (AM) and their prompt integration with the wide range of clinical and basic research initiatives on campus. To meet these requirements, the department faculty are organized in two transparent layers: ‘Integrators’, who will have joint appointments with the Department of Bioinformatics and one of the clinical or basic science departments at UTSW; and ‘Innovators’, who will build the CS and AM foundation for the campus. The vision of this organization relies on the following premises:
The primary bioinformatics applications on campus relate to: i) genomics, ii) multi-scale imaging, iii) understanding disease and therapeutic intervention in networks of multi-factorial processes (systems biology and systems pharmacology), and iv) clinical decision making based on multi-model and multi-parametric information. The Department recruits investigators as ‘Integrators’ who bring to the campus expertise in genomic data modeling, including evolutionary biology, both at the population- and single-cell levels, in regulatory network analysis, in bioimage informatics both at the single-cell and tissue (brain) levels, and in digital pathology and medical informatics. There is significant overlap between these areas. For instance, genomic data modeling is in many aspects related to the inference of regulatory networks; image based-modeling of cell and tissue morphogenesis involves modeling of complex mechano-chemical signaling networks; and much of digital pathology can be cast as a problem of probing tissue morphology. Therefore, despite topical diversity, the Department exists as a unit with synergy among the ‘Integrator’ faculty.
Advances in application-driven bioinformatics depend on steady innovation in Computer Science and Applied Mathematics. The ‘Integrator’ faculty are complemented by investigators who focus on foundational aspects of CS and AM as pertain to bioinformatics and data science at large. Without the core of ‘Innovators’ the Department will lack intellectual renewal and is at a risk of losing relevance over a short time period. Cross-fertilization between ‘Integrator’ and ‘Innovator’ programs are fostered by careful team-oriented recruitment, an open space design, and shared infrastructure.
Topically diverse bioinformatics applications build on common foundations in Computer Science and Applied Mathematics. The two layers of the Department structure are tightly connected through a web of scientific synergy. Although our recruitment strategy is flexible, ‘Integrator’ faculty tend to run more established programs, while ‘Innovator’ faculty are recruited from a pool of early-stage investigators who bring cutting-edge CS and AM research to the campus. A key component of this senior-junior mix is that the ‘Integrator’ faculty will be responsible for the recruitment and mentoring of junior ‘Innovator’ faculty. This ensures that the fundamental CS and AM questions addressed by the Department are anchored within the biomedical research on campus.