Developing a successful open-informatics endeavor for DOE systems biology required key input and skills from several groups within the scientific community. Broadly these groups represented plant and microbial researchers who design experiments and generate data; computational biologists and bioinformaticians who will interpret and simulate data; and computer scientists, database developers, and software engineers who will develop Kbase infrastructure. Representatives from these communities participated in the five Kbase workshops. In addition to contributing to this implementation plan, workshop participants also addressed the cultural transition the informatics community will need to make from individual project-based efforts toward research community-based informatics.
Co-organizers: Folker Meyer, Argonne National Laboratory (ANL); Susan Gregurick, U.S. Department of Energy (DOE); Peg Folta, Lawrence Livermore National Laboratory; Bob Cottingham, Oak Ridge National Laboratory (ORNL); and Elizabeth Glass, ANL
This Kbase workshop focused on applications of cloud computing. It brought together researchers in the computing, systems biology, bioinformatics, and computational biology fields. Modern genomics studies use many high-throughput instruments that generate prodigious amounts of data. For example, a single run on a current sequencing instrument generates 30-40 gigabytes of sequence data. The situation is complicated further by the democratization of sequencing; many small centers now can independently create large sequence datasets. Moreover, the immense amount and variety of omics data that must be integrated with genomics data to model and study organisms at a systems level create unique opportunities in computational biology. Consequently, the rate of sequence and related data production is growing faster than our ability to analyze these data.
Cloud computing provides an appealing possibility for on-demand access to computing resources. Many computations can be considered embarrassingly parallel and should be ideally suited for cloud computing. However, challenging issues remain, including data transfer and local data availability on the cloud nodes. In discussing the feasibility of using cloud computing for Kbase, clear needs included flexible architecture and input/output (I/O), high-quality reference data and standards, and prioritized workflows.
Co-organizers: Catherine Ronning, DOE; Susan Gregurick, DOE; Ed Kaleikau, USDA; Gera Jochum, USDA; and Bob Cottingham, ORNL
This workshop was jointly convened by DOE BER and the U.S. Department of Agriculture National Institute of Food and Agriculture. It brought together 100 plant scientists, geneticists, breeders, and bioinformatic specialists to discuss current issues facing plant breeders in light of ever-increasing amounts of genomic data. The workshop featured lectures by leaders in the plant breeding, genomics, and bioinformatics communities. These presentations set the stage for afternoon breakout discussions by addressing the data needs of more-applied breeding programs and describing resources emanating from more-fundamental plant genomics and bioinformatics research. The overarching question was, "How can we best design the Knowledgebase to have the flexibility to grow with and adapt to new data and information challenges in the future?"
A key objective was to specifically identify the requirements for effectively developing data capabilities for systems biology as applied to plants, particularly the research and development of plant feedstocks for biofuels. The current state of plant informatics is represented by many disparate databases primarily focusing on specific taxonomic groups or processes. To enable a systems biology approach to plant research, integrating all types of data (including molecular, morphological, and omics) for bioenergy-relevant plant species is important. Thus, a challenge for Kbase will be to develop uniformity of data format and database architectures to effectively integrate diverse data types and enable user-friendly acquisition and analysis.
Co-organizers: Susan Gregurick, DOE, and Bob Cottingham, ORNL. Co-chairs: Adam Arkin, Lawrence Berkeley National Laboratory (LBNL), and Robert Kelly, North Carolina State University]
Workshop participants discussed the current, near-, and long-term prospects for microbial systems biology research in the context of the Knowledgebase. The rapidity with which new genome sequence information appears in public databases is presenting a growing challenge for the data storage, analysis, and utilization necessary to foster scientific and technological advances. The systems biology framework has arisen in response to this challenge, but new computing strategies are needed to take advantage of this new context for examining microbial biology. The "monoculture" paradigm has been quite productive and will continue to be at the heart of microbiology. However, monocultures are not representative of how microbial systems exist in nature. To this end, metagenomics has provided a means for examining microbial complexity, but complementary functional information is still needed to understand the "metaphenotype."
In biology, a grand challenge is to predict phenotype from genotype. This challenge is complicated in microbes because a significant fraction of microbial genomes interacts with other organisms and not all genes are continuously expressed. The scientific community is relatively well developed in terms of measuring various types of omics data, but challenges remain for highly complex environments, such as soil and sediments. In the long term, Kbase will be faced with capturing and interrelating data about all these processes at scales from molecules to meters. Several workflows were initiated at this workshop that have been further refined and incorporated in this implementation plan. These include Microbial Scientific Objective 1: Reconstruct and Predict Metabolic Network to Manipulate Microbial Function and Microbial Scientific Objective 2: Define Microbial Gene Expression Regulatory Networks.
Co-organizers: Susan Gregurick, DOE, and Bob Cottingham, ORNL. Co-chairs: Victor Markowitz, DOE JGI and LBNL, and Jill Banfield, University of California, Berkeley
The focus of this Kbase workshop was to discuss scientific objectives and challenges for data handling and knowledge integration specific to the study of microbial communities or metagenomes. Some topics also were pertinent to all development and initial implementation of knowledgebases for the broader biological community. A main workshop theme was to discuss Kbase as a project that would build on existing systems for managing and analyzing omics data while achieving a higher level of support for the scientific community. Several objectives and workflows were initiated at this meeting.
Co-organizers: Susan Gregurick, DOE; Bob Cottingham, ORNL; and Brian Davison, ORNL
The focus of this workshop was to define detailed requirements for initial priorities, a robust design, and implementation plans to create Kbase, this workshop involved 80 participants representing university, national laboratory, and international scientists, as well as key stakeholders (plant and microbial genomic researchers, bioinformaticians, computer scientists, database developers, and software engineers). Workshop participants also included representatives from the DOE JGI; DOE's Bioenergy Research Centers; the National Science Foundation's (NSF) iPlant; and the National Institutes of Health's (NIH) National Cancer Institute and National Center for Biotechnology Information.
Emphasis was placed on prioritizing clear scientific objectives and specifying the associated tasks and requirements for achieving these objectives. Participants were charged with developing and prioritizing three to five scientific objectives in three areas: microbial, metacommunity, and plant research. Extensive pre-meeting conference calls helped lay the groundwork for workshop participants to develop scientific requirements, time frames, and the level of effort expected for Kbase support of each objective. Once finalized, the requirements were translated into implementation plans for each objective. Workshop discussions also addressed system architecture and governance for the initial system, however, participants were not charged with defining funding or contractual structures. A consensus among participants was that initial Kbase efforts cannot be all things for all users. Showing strong success in a few areas is better than making minimal progress in many areas. Workshop participants also expressed continued support for Kbase principles identified at previous workshops:
(1) science drives Kbase development;
(2) the project should be a community effort;
(3) Kbase should support open access and open contribution; and
(4) Kbase resources and capabilities should be distributed.
In addition to defining scientific objectives, the systems biology community also articulated the need to define research workflows that enable scientists to compare and contrast different methods. This was deemed a necessary component of the implementation plan, because workflows will form a basis for researcher interactions within the Kbase.
This work, was sponsored by BER in the DOE Office of Science with American Recovery and Reinvestment Act (ARRA) 2009 funding and was performed at Oak Ridge National Laboratory (ORNL). ORNL is managed by UT-Battelle, LLC, for DOE under contract DE-AC05-00OR22725.