Vendor Workshops
Artificial Intelligence & Data Mining
O.R. in Biology and Medicine (sponsored by NSF)

Saturday, November 4, 8:00am-5:00pm (lunch included)
Partially funded by National Science Foundation
Planning Committee:  Harvey Greenberg and Ariela Sofer (co-Chairs), Steve Billups, Bala Krishnamoorthy, Cesar Rego, Wilbert Wilhelm

WORKSHOP SUMMARY
In recent years the frontiers of biological and medical research are becoming increasingly dependent on sophisticated modeling, analysis, and computational techniques. Operations research and mathematics are rapidly emerging as vital tools for investigating complex biological systems. The interaction between the fields of OR and biomedicine has been fruitful and synergistic. Exciting advances in biomedical research have been made possible by new mathematical and computational techniques. In turn, the biomedical research has spawned new mathematical developments in operations research. 

This interaction of these disciplines, however, faces many challenges. As mathematics becomes increasingly central to biology, it is crucial that biologists have understanding of the capabilities and limitations of these techniques. It is similarly important that the math/OR researcher develop a sufficient working knowledge of biological concepts and processes. The purpose of this workshop is to bring together researchers in biology, medicine, operations research, mathematics, statistics, computer science and other disciplines, for the purposes of examining the challenges of this cross-disciplinary research. Specifically the workshop will address the following issues:

• Research - what opportunities are there for OR to solve problems in biology? What are the opportunities for biologists in collaborating with OR researchers? How does one establish collaborations?
  
  
• Education - how do we provide MS and/or PhD students with the knowledge and skills to conduct interdisciplinary research at the interface between the biomedical sciences and the mathematical and OR disciplines? Should we be training students for computational biology at the MS level? Is the industrial job market sufficiently mature so these graduates could get jobs?

• Academic Careers - how do we encourage the professional development of faculty in developing educational approaches appropriate to an interdisciplinary biology/OR education? How does one best integrate teaching and research activities across traditional disciplinary departments? What are the mechanisms for eliminating institutional, departmental, and administrative barriers to interdisciplinary research and education?
  
  
• Funding - cross-disciplinary projects often find it difficult to compete with applications submitted by the different individual disciplines, and it is often difficult to choose which discipline to represent as applicant for funds. What funding opportunities exist?

PROGRAM
8:15am-9:30am
Forum meeting – goals and strategies (continental breakfast)

9:30am-10:30am
Graduate Education at the Interface of the Computational and Biological Sciences
Daniel Voytas, Department of Genetics, Development and Cell Biology, Director of Ph.D. Program in Bioinformatics and Computational Biology, Iowa State University

Computational biology is a new field that uses genomic sequence, expression and structural data to understand complex biological pathways, processes and systems. Practitioners of computational biology must be competent not only in the biological sciences, but also in computer science, math or statistics. Achieving this breadth of expertise presents challenges for graduate education and requires new approaches that fall outside the paradigm of traditional, discipline-based graduate training.

10:30am-11:00am
Break

11:00am-12:00pm
Models of the Physician's Problem in Treatment Planning: The Practitioner's Dilemma
Mark Langer, MD, Clinical Radiation Oncology, School of Medicine, Indiana University

Modern radiotherapy planning for each patient involves the assignment of values to thousands of delivery components that determine the distribution of dose within the body, and ultimately the chances of cure and complication. Dose, like temperature, can be mapped to each point in the body, but it is the distribution of dose throughout an organ that determines the biologic effect. In the many organs, preservation of physiologic function demands that some fraction of an organ mass be held below a threshold level for damage. Finding sacrificed and spare elements of each organ to satisfy the tolerance rules and admit the largest possible dose into the tumor forms a large-scale problem of combinatorial complexity. Recent advances in the mechanics of delivery makes it possible to image the tumor and surrounding structures over the course of treatment, demanding planning in response to observed or predicted changes in geometry over time.

Responding to changes in tumor over time requires that both cumulative dose constraints, and constraints on each part of the treatment course be observed, giving rise to adaptive, model predictive, and robust model predictive control problems. Effective operations demands that the physician, physicist and dosimetrist communicate objectives and constraints, provide procedures to examine sensitivity to uncertainty in the specifications or errors in delivery, adapt to observed or predicted change in anatomy over treatment, and develop a strategy for constructing serial relaxations of constraints in the not uncommon event that the initial directives are too overreaching to be physically feasible. In the near future, it is expected that biologic functional imaging of cellular metabolism will drive the scheduling and mechanics of many kinds of tumoricidal agents, generalizing the treatment-planning problem to all of oncology. Planning problems will be posed, and the need for computational and operational solutions described.

12:00pm-1:30pm
Lunch + discussions

1:30pm-2:30pm
Frontier Opportunities for Funding in Biology and Medicine
Matthew Realff, Program Director, Design, Manufacture, and Industrial Innovation Division, National Science Foundation

There has been an explosion in data being generated in biology and medicine at many different scales. High throughput screening, genome sequencing, and multi-genome sequencing have shed light on the molecular mechanisms of evolution, disease and for biotechnology applications. Satellite imagery and large-scale ocean and ecosystem observatories are doing the same for landscape level processes such as migration, specie distribution and disease spread. These new streams of data require new models and solution methodologies to extract useful information and to plan experiments, as well as enabling decision-making for control of epidemics, biotechnology, and engineering systems inspired by biological processes. This talk will review opportunities for the discipline of operations research to make contributions to the areas of biology and medicine, and focus specifically on funding opportunities available at federal agencies with a particular focus on the National Science Foundation.

2:30pm-3:00pm
Break

3:00pm-4:00pm
Grant Programs of the National Library of Medicine (NLM)
Milton Corn, Director, Division of Extramural Programs, National Library of Medicine

NLM has supported research in application of computing to biomedicine for over 20 years. The major theme of the awards in the early years of the programs was clinical informatics and the information management of health care delivery. Although a significant portion of NLM’s grant portfolio is now dedicated to bioinformatics and translational informatics, NLM continues to have a strong interest in the information of health care, and in the related fields of decision support, multi-media, information retrieval, and use of healthcare information for health services research and public health research.

At present, grants are available for fundamental and applied research in biomedical informatics and bioinformatics. Areas of research interest include: healthcare delivery, representation, organization and retrieval of biomedical and biological data and images; enhancement of human intellectual capacities through virtual reality, dynamic modeling, artificial intelligence and machine learning; medical decision-making; linguistic analyses for natural language processing and understanding; informatics topics relevant to public health and informatics for disaster management.

4:00pm-5:30pm
Forum discussions

CONFERENCE BACKGROUND

Summary of CCB Roundtable Discussion 1:
OR and the New Biomedical Research Profession, available at
www-math.cudenver.edu/~hgreenbe/CCB_RT1summary.pdf

Questions that were addressed:

1. What are opportunities for OR to solve problems in biology?
2. How should we change OR education to prepare students for working in this field?
3. What funding is available, and how might we consort?

During the meeting, “biology” was extended to “biology and medicine.”