Preliminary CBCB Planning Document

CONSORTIUM FOR BIOINFORMATICS
&
COMPUTATIONAL BIOLOGY

A proposal coordinated
by
H. Ted Davis and Steve Cawley

Introduction.

The deciphering of the human genome, as well as of those of numerous other animal, plant, and microbial species, is just the opening act of major advances in science and technology. The 21st century will see tremendous advances in understanding the basis of life, the creation of healthier and hardier foods, the development of safer and more specific drugs, and the curing or successful treatment of inherited and acquired diseases. These advances will be built on the genome work, but will require much additional research. Deciphering the genome means identifying the DNA (DeoxyriboNucleic Acid) sequences that provide the genes of a species. These sequences by themselves form massive databases that require information science for storage, organization, indexing, and mining techniques. But beyond that, there are additional, even larger databases, that bioinformatics has evolved to handle. The genetic DNA governs the production of the various proteins that are the building blocks of life. How the proteins are synthesized, how they interact with each other, and how the various complicated feedback loops in biological organisms interact are all major challenges that biomedical science has to and seeks to elucidate. In order to carry out the thorough systems engineering analysis that is required to provide major health and economic advances for society, we will require better bioinformatics and computational biology expertise. It is only by the use of modern digital technology techniques that the complexities of the problem can be handled.

Advances in biology in the up-coming decades are going to be driven by those universities that develop broad, deep, interdisciplinary, and coherent programs in bioinformatics and computational biology. Being a comprehensive university with colleges of biology, the physical sciences and engineering, and the agricultural and health sciences, the University of Minnesota has the pieces that, if properly coordinated, could put the University in a commanding position in the field. The purpose of this document is to propose to use the resources from the Microsoft class action settlement to create a consortium for bioinformatics and computational biology that will make Minnesota a major power in the field. Minnesota was passed by in the first major wave of molecular biology because of lack of commitment and vision on the part of the University’s leadership. Thanks to Mark Yudof’s initiatives in the biological sciences and digital technology, we have the opportunity to rejoin the biological revolution, if we bring together the right university groups and focus our resources.

Making Our Case.

Optimal progress in life sciences research will not be possible without computational biology/informatics expertise. Therefore, the university’s substantial investments in molecular and cellular biology and in genomics and proteomics will not be fully realized without comparable enhancement of our computational biology/bioinformatics abilities. Although units of the university have invested in hardware and software to support this research, we have not yet adequately developed the academic expertise in computational biology/bioinformatics appropriate for our institution. Indeed, on two occasions in the past two years the University of Minnesota has been unable to even assemble a competitive application to become one of several national centers for bioinformatics funded by the NIH because of insufficient expertise and leadership recognition in bioinformatics. We have several faculty members who are renowned for their individual work, but no visionary computational biologists who can lead a University-wide program. In no other area related to life sciences research are we so obviously deficient.

Computational biology/bioinformatics is clearly an interdisciplinary activity involving computer scientists, mathematicians and biologists. Computational biology/bioinformatics currently does not comfortably exist in any one college or school. A campus-wide Ad hoc Bioinformatics/Computational Biology Advisory Group convened October, 2003 identified the need for a single senior leader for bioinformatics who could help to develop the overall vision for bioinformatics at the university and bring together the numerous, currently separate components of our computational biology/bioinformatics activities. It was agreed that recruiting a national leader in computational biology/bioinformatics was the most likely route to success, but no further progress was made because, even collectively, the colleges and other academic units could not assemble an appropriate recruitment package.

With the Microsoft settlement, we now have an opportunity to meet a critical research need in life sciences research and education by recruiting a leader and additional junior faculty in the interdisciplinary field of computational biology/bioinformatics. Such a recruitment into a consortium for bioinformatics and computational biology will catalyze the consolidation of separate computational biology/bioinformatics entities (e.g., the Computational Genomics Laboratory of the Supercomputing Institute and the AHC Center for Computational Genomics and Bioinformatics) and enable us to make other strategic hires to establish the U of M as a national leader in the area.

Organizational Structure.

The Consortium for Bioinformatics and Computational Biology (CBCB) would be headquartered in the Digital Technology Center and operated for the benefit of all the collegiate units of University of Minnesota, but would be managed by the cooperating colleges which put up matching resources for the CBCB. Members of the CBCB and its supporting hardware and software would be distributed between the Minneapolis and St. Paul campuses. Members of the CBCB would include faculty and P&A personnel already active in bioinformatics and computational biology and a director, new faculty and P&A personnel to be added to achieve the balance and strengths needed to form a world-class center.

The CBCB director will report to the director of the DTC, whose policies with regard to CBCB activity will be reviewed by a Board of Directors appointed by the cooperating colleges and will include the Director of the DTC. Another board, the Planning Committee, will be appointed by the cooperating colleges to map out the best strategy for building the consortium and optimizing the University’s position in bioinformatics and computational biology. The cooperating colleges would include the Medical School, the College of Biological Sciences, the Institute of Technology, the College of Agriculture, Foods, and Environmental Sciences, the School of Public Health, the College of Veterinary Medicine, the School of Pharmacy, the School of Dentistry, and the College of Natural Resources.

The Plan.

We propose that the Consortium for Bioinformatics and Computational Biology be formed in the DTC using the $2.5M cash award and the $2.5M voucher award from Microsoft as seed money and computer resources that would establish a world-class center that would be fully funded by the cooperating colleges after the first five years. There are at least seven major colleges that have a strong interest in the areas represented by the consortium and that would benefit greatly by its creation.

There are activities in bioinformatics and computational biology already going on in many colleges of the University. However, they are not fully effective, because there is duplication of effort, lack of coordinating leadership, lack of well-planned service facilities, and some weaknesses in strategic areas (in particular, theoretical biology). We propose that the CBCB become the focus of the University’s data handling and computational services for bioinformatics and computational biology, as well as provide leadership in theoretical biology and the computational science of biology.

We propose to use the Microsoft cash at a rate of $500K per year. $200K per year would be used to hire in the CBCB 2 high level P&A researchers versed in the biology as well as the computational science relevant to bioinformatics and computational biology. Although the DTC already serves some of the research activity in bioinformatics and computational biology (through a new computer lab opened in the Cargill building on the St. Paul Campus, the Minnesota Supercomputing Institute, and funds furnished to the genomics center in the Academic Health Center), the strengths of DTC lie more in computer science and scientific computation. The 2 new P&A hires are designed to deepen the biological competence of the CBCB. It is also proposed that the resources currently available for genomics and proteomics in the Academic Health Center be folded into the CBCB. After five years, the cooperating colleges must agree to continue devoting $200K per year for the P&A personnel. Another $100K per year will be invested in the University Library as described below.

The proposal for the remaining $200K per year of the Microsoft cash is to fund at $50K per year 4 entry-level faculty positions to strengthen our position in bioinformatics and computational biology. These would be cluster hires in the sense that any college can try to make a strategic hire, and, if approved by the CBCB Planning Committee, can make the hire if the college picks up the rest of the salary of the new faculty member. After five years, the colleges getting these new faculty will have to pick up their whole salary.

The cooperating colleges will be asked to collectively fund a senior faculty position (salary and start-up costs) to hire as soon as possible a star in theoretical biology (i.e., in computational biology/bioinformatics) as Director of the CBCB. The formula for each cooperating colleges contribution to this hire will be proposed to the Board of Directors by the Planning Committee.

We propose to use the $2.5 M in vouchers to purchase Linux and Apple clusters for bioinformatics and computational biology and provide system administration support for them. One of the clusters would likely be a Paracel Cyclone system that includes the Paracel GeneMatcher2 high-speed, massively parallel supercomputer specially designed for genomic analysis. There would also be some specialized software purchases, and high-end storage systems to handle the large data sets that arise in this research. The cost of the hardware and software would be about $1.5 M, and would be heavily front-loaded. $100K per year of the cash settlement and the other $1 M of vouchers would be used at $300 K per year to hire two system administrators for the clusters and fund the library at $100K per year to develop and administer a repository of research results and data that would use emerging “institutional repository” software (such as DSpace from MIT) to host, provide access to, and archive the output of the Consortium and related research at the University. The way the vouchers would be converted to cash would be to swap vouchers for cash with the IT Computer Laboratories, which spends much more than that each year on desktop computers and workstations. After five years, the $200 K per year cost of the two system administrators would be funded through grants raised by participants in the consortium and the $100K per year for the library would be covered by the University.

The exact configurations of the hardware to be obtained would be decided by the Planning Committee.