BMRB

Biological Magnetic Resonance Data Bank


A Repository for Data from NMR Spectroscopy on Proteins, Peptides, Nucleic Acids, and other Biomolecules
Member of WWPDB

BMRB Aims and Policies

Introduction

BioMagResBank (BMRB) is the publicly-accessible depository for NMR results from peptides, proteins, and nucleic acids recognized by the International Society of Magnetic Resonance and by the IUPAC-IUBMB-IUPAB Inter-Union Task Group on the Standardization of Data Bases of Protein and Nucleic Acid Structures Determined by NMR Spectroscopy. In addition, BMRB provides reference information and maintains a collection of NMR pulse sequences and computer software for biomolecular NMR. Access to data in BMRB is free directly from its web site (URL http://www.bmrb.wisc.edu) and ftp site (ftp.bmrb.wisc.edu) and will remain so as public funding permits. The concept of a biomolecular NMR data bank was developed under a five-year research grant awarded to the University of Wisconsin-Madison from the National Library of Medicine, National Institutes of Health. BMRB at the University of Wisconsin-Madison has been supported since 1 September 1996 by the National Library of Medicine.

Aims

In collaboration with the Protein Data Bank (PDB, Rutgers University) and Nucleic Acid Data Bank (NDB, Rutgers University), BMRB aims to develop into the collection site for structural NMR data in proteins and nucleic acids. Steps are being taken at BMRB to achieve this capability, which will include transmission of all relevant data to the PDB and NDB collections. In addition, BMRB has the goal of archiving NMR-specific data including assigned chemical shifts, J-couplings, relaxation rates, and chemical information derived from NMR investigations (for example, hydrogen exchange rates and pKa values). In developing these collections BMRB intends to be responsive to the needs and priorities of the scientific community. The operating policy at BMRB is monitored and shaped by its Advisory Board which meets once each year to review progress and set priorities. The Advisory Board is composed of representatives from laboratories that originate and/or use data within the BMRB purview.

Author-initiated data submission

Although BMRB contains data that have been entered by its staff members from the literature, its limited budget mandates that data must be contributed directly by scientists who are data producers. This policy was articulated in BMRB's grant proposal. Thus BMRB's major thrust has been (1) to develop a comprehensive format for biomolecular NMR data that will serve the present and future needs of the biomolecular NMR community, and (2) to develop and test software tools for more convenient data deposition by authors.

Data format (flat-file and data exchange)

After extensive review of current alternatives, the Self-defining Text Archival and Retrieval (STAR) format developed by Hall and Spadaccini was chosen as the foundation for the BMRB flat-file and data-exchange format. The NMR version of this, NMR-STAR, is embodied in the BMRB Data Deposition Form which is available from the BMRB web site. Other groups are beginning to provide NMR-STAR as an output format, for example, the authors of AQUA (Utrecht University, The Netherlands) and Felix (MSI, San Diego CA, USA). All data at BMRB are available in the NMR-STAR format, and several scientific groups around the world have begun to deposit their data directly as NMR-STAR files. These are validated by the BMRB staff and are added to the collection available from the BMRB web site.

Tools to assist with data deposition

Several tools have been developed at BMRB to assist depositors in creating valid files for deposition. Data table generators for depositing chemical shifts of proteins and nucleic acids in the NMR-STAR format have been released and are available from the BMRB web site. Software graciously provided by PDB has been adapted and extended to support author input of NMR chemical shift data for proteins and nucleic acids. This software, ADIT-NMR is the preferred method for depositing NMR data and has been in use since March of 2004.

Data accepted

BMRB will accept: quantitative NMR data for a peptide, protein, nucleic acid, or polysaccharide and their ligands and cofactors assigned on an atom-specific basis, and data derived from the analysis of such data. Submitted data are to be in valid NMR-STAR format; they may be created by tools that output data in this format. A given BMRB submission contains data related to a single biomolecular system from a single publication (data on the same system from multiple publications, or from multiple systems from a single publication, should be broken into separate submissions). A BMRB submission can contain data that have not been published or for which no publication is intended by the author(s). The types of NMR data that can be included in a deposition and the mandatory supporting information accompanying a deposition are reviewed yearly by the BMRB Advisory Board.

Accession numbers

Journals typically recommend (or require) that articles containing structural NMR data provide the accession number under which data in machine-readable form can be found. Upon receipt of a new data set that contains all mandatory supporting information, BMRB will send an accession number to be included in the publication. If following publication, a reader cannot find a data set at BMRB that corresponds to the published accession number, they are requested to notify both the journal and BMRB.

Procedures for processing and releasing data sets

Receipt of data sets is announced on the "Entries Being Processed" page of the BMRB web site. This listing consists of the BMRB accession number, date of receipt, and additional notes that track its status, e.g, "returned to author" for additional information or verification. A protected site is available where depositors can inspect their entries and make them available for view by reviewers of manuscripts. Completed files will be placed "on hold" if authors so request; the hold normally is until publication. In exceptional cases, authors may request that BMRB withhold the release of data on a particular study for a time up to one year following deposition. After final processing by BMRB, new submissions are merged with others accessible from the BMRB data retrieval systems.

Validation process

Submissions are checked for completeness (mandatory data), correct syntax, and internal consistency. Potential typographical errors and outliers in the data are brought to the attention of the corresponding author. The author has the option of making corrections or declining suggestions. Before a file is released, BMRB staff members add and verify links to other databases.

Data export

All data currently are exported in NMR-STAR format and in ASCII files exported as relational database tables. All exported data have the author(s)' name(s) attached to the entry and list the primary publication associated with the BMRB file. Users are requested to acknowledge BMRB in publications: a sample acknowledgment is available from the BMRB web site.

Correction or revision of entries

Authors are encouraged to update their BMRB files to take new results into account. Entries may be deleted (from the actively-accessed database) at the author's request. The BMRB staff may modify a file to update the citation or to upgrade it to maintain consistency with the data format. As entries are updated, a statement describing the revision, the date, and the author of the revision will be added to the revision history loop located near the beginning of the file.

NMR pulse sequences and software

BMRB maintains a library of NMR pulse sequences and computer software for biomolecular NMR. Submission instructions are available from the BMRB web site. Files are supplied directly as submitted, with no guarantee or recommendation as to their usefulness. The authors who submitted the files are listed, and they should be contacted for further information or to report difficulties.

Criticisms and suggestions

The BMRB team and its Advisory Board welcome feedback from data depositors and users. These may be addressed to:
John L. Markley, Head BMRB, markley@nmrfam.wisc.edu
Eldon L. Ulrich, Director BMRB, elu@bmrb.wisc.edu