The Blueprint Initiative Samuel Lunenfeld Research Institute (c)2005 The Blueprint Initiative of Mount Sinai Hospital. All Rights Reserved. README: BLUEPRINT DATA FTP AREA ftp://ftp.bind.ca/pub/BIND/data/bindflatfiles/bindindex/ LAST UPDATE OF THIS README: February 3, 2005 This area contains selected fields for BIND records in plain-text, tab-delimited format. Files are released daily and are cumulative to the date specified in the file name; i.e., yyyymmdd. These files may be used to locate BIND records that describe a given biomolecule. BIND records contain many other fields than those listed in these indices. Complete BIND records are available in XML or ASN.1 format from ftp://ftp.bind.ca/pub/BIND/data. A complete description of these BIND index files is given below. DIRECTORY----------------------------------------------------------------------- yyyymmdd.ints.txt Contains selected fields for all BIND interactions. yyyymmdd.refs.txt Contains publication references and experimental methods that are relevant to BIND interactions. yyyymmdd.nrints.txt Contains selected fields for a non-redundant list of interactions in BIND. yyyymmdd.taxon.txt Taxon identifiers and corresponding scientific names for organisms described in BIND. yyyymmdd.complexes.txt BIND complexes index. This table lists all complex records in BIND along with their titles and the number of subunits in the complex. yyyymmdd.complex2subunits.txt BIND complex subunits index. This table lists the subunits present in each BIND complex. yyyymmdd.complex2ints.txt BIND complex interactions index. This table lists interaction records from BIND that are part of the complex. yyyymmdd.complex2matrix.txt BIND complex MATRIX representation. This table lists all potential interactions that may be occurring in a BIND complex when the topology of the complex is unknown. yyyymmdd.complex2spoke.txt BIND complex SPOKE representation. This table lists interactions between a central subunit (used to detect members of a complex; for example, the subunit used as 'bait' in an immunoprecipitation) and all other subunits in that complex when the topology of the complex is unknown. yyyymmdd.complexes2refs.txt BIND complexes refs index. This file lists publication references that support complex records. yyyymmdd.citations.txt Citations for BIND record references listed in this index - see also refs.txt and complexes2refs.txt yyyymmdd.nr_bind_proteins.fsa This file lists all of the non-redundant protein sequences from BIND in FASTA format. The definition line uses the sequence record GI and accession for the best-annotated sequence in the group of redundant proteins. See yyyymmdd.redundant_gi.txt in this README for more details. yyyymmdd.redundant_gi.txt BIND redundant GI's index. yyyymmdd.labels.txt This file lists the short labels attached to each interactor object for a given interaction ID. ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND interaction index (yyyymmdd.ints.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- rgid INTEGER redundant group id 1 bid INTEGER bind record id 2 A-type text molecule A type 3 A-db text molecule A database 4 A-acc VARCHAR molecule A accession 5,6 A-id INTEGER molecule A id 5,7 A-tax INTEGER molecule A taxon-id 8 B-type text molecule B type 3 B-db text molecule B database 4 B-acc VARCHAR molecule B accession 5,6 B-id INTEGER molecule B id 5,7 B-tax INTEGER molecule B taxon-id 8 AB text same AB order in both 9 rgid and iid? ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND interaction references and methods index (yyyymmdd.refs.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- rgid INTEGER redundant group id 1 bid INTEGER BIND record identifier 2 pmid INTEGER PubMed Identifier 10 method text method 11 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- Non-redundant BIND interaction index (yyyymmdd.nrints.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- rgid INTEGER redundant group identifier 1 A-type text molecule A type 3 A-db text molecule A database 4 A-acc VARCHAR molecule A accession 5,6,12 A-id INTEGER molecule A id 5,7,12 A-tax INTEGER molecule A taxon-id 8 B-type text molecule B type 3 B-db text molecule B database 4 B-acc VARCHAR molecule B accession 5,6,12 B-id INTEGER molecule B id 5,7,12 B-tax INTEGER molecule B taxon-id 8 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- Taxon identifiers and organism names (yyyymmdd.taxon.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- taxon INTEGER taxon identifier 13 name text taxon name ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND complexes index (yyyymmdd.complexes.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- bcid INTEGER BIND complex record identifier 14 numsubs INTEGER number of subunits 15 descr text description of complex 16 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND complex subunits index (yyyymmdd.complex2subunits.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- bcid INTEGER BIND complex record identifier 17 type text molecule type 3 db text molecule database 4 acc VARCHAR molecule accession 5,6 id INTEGER molecule id 5,7 tax INTEGER molecule taxon-id 8 short-label VARCHAR short label of molecule 22 other-names VARCHAR list of synonyms 23 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND complex interactions index (yyyymmdd.complex2ints.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- bcid INTEGER BIND complex record identifier 18 bid INTEGER BIND interaction record identifier ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND complex MATRIX representation (yyyymmdd.complex2matrix.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- bcid INTEGER BIND complex record identifier 19 A-type text molecule A type 3 A-db text molecule A database 4 A-acc VARCHAR molecule A accession 5,6 A-id INTEGER molecule A id 5,7 A-tax INTEGER molecule A taxon-id 8 B-type text molecule B type 3 B-db text molecule B database 4 B-acc VARCHAR molecule B accession 5,6 B-id INTEGER molecule B id 5,7 B-tax INTEGER molecule B taxon-id 8 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND complex SPOKE representation (yyyymmdd.complex2spoke.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- bcid INTEGER BIND complex record identifier 20 A-type text molecule A type 3 A-db text molecule A database 4 A-acc VARCHAR molecule A accession 5,6 A-id INTEGER molecule A id 5,7 A-tax INTEGER molecule A taxon-id 8 B-type text molecule B type 3 B-db text molecule B database 4 B-acc VARCHAR molecule B accession 5,6 B-id INTEGER molecule B id 5,7 B-tax INTEGER molecule B taxon-id 8 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND complexes refs index (yyyymmdd.complexes2refs.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- bcid INTEGER BIND complex record identifier 21 pmid INTEGER PMID references 10 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND citations index (yyyymmdd.citations.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- pmid INTEGER PubMed Identifier 10 citation text authors, title, journal, pages 24 ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND redundant GI's index (yyyymmdd.redundant_gi.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- best annotated gi INTEGER GenBank identifier 25 gi INTEGER GenBank identifier ================================================================================ ================================================================================ -------------------------------------------------------------------------------- BIND labels index (yyyymmdd.labels.txt) -------------------------------------------------------------------------------- column name type description note -------------------------------------------------------------------------------- bid INTEGER bind record id 2 A-short-label VARCHAR short label of molecule A 22 B-short-label VARCHAR short label of molecule B 22 A-other-names VARCHAR "|" delimited list of short-label synonyms for molecule A 23 B-other-names VARCHAR "|" delimited list of short-label synonyms for molecule B 23 ================================================================================ ================================================================================ ================================================================================ notes: 1. This number (rgid) is used to group together a number of BIND records that describe interactions between the same two molecules (although they may be referred to by different identifiers). As such, the 'rgid' corresponds to a group of BIND record identifiers. See the table 'yyyymmdd.nrints.txt'. The rgid is unstable and may not always refer to the same group of interactions from one release of this index to another. Note that two small molecules with identical database names and identifiers are considered to be the same even if they came from different organisms. For example, see BIND interaction 232 and 233. These interactions are redundant. 2. Every BIND record is represented in this table at least once if it describes an interaction between two objects that are of any type besides "not-specified". Records where one of the two objects is of the type "not-specified" are commonly used to describe complexes that have an unknown topology. These interaction records will not be found in this index (for example, BIND interaction 6558). 3. Fields named 'A-type', 'B-type' or 'type' describe the molecule type and are a choice of: not-specified protein DNA RNA small-molecule complex gene photon The text is listed in this column (for example; "DNA" or "protein") For more information see the BIND specification: BIND-object->BIND-object-type-id. 4. Fields named 'A-db' , 'B-db' or 'db' describe the database where the record for the molecule is found. This is a choice of: GenBank: always the case for DNA, RNA and protein type molecules; also used for type 'gene' KLOTHO: Biochemical Compounds Declarative Database (http://www.biocheminfo.org/klotho/) may be used for molecules of type 'small-molecule' LIGAND: Database of Chemical Compounds and Reactions in Biological Pathways (http://www.genome.ad.jp/ligand/) may be used for molecules of type 'small-molecule' MMDB: Molecular Modelling Database (http://www.ncbi.nlm.nih.gov/Structure/MMDB/mmdb.shtml) may be used for molecules of type 'small-molecule') BIND: Biomolecular Interaction Network Database (http://bind.ca/) always used for molecules of type 'complex' CHEMICAL ABSTRACTS SERVICE: (http://www.cas.org/) may be used for molecules of type 'small-molecule' The text name of the database is listed in this column (for example; GenBank or MMDB) For more information see the BIND specification: BIND-object->BIND-object-type-id->BIND-id and BIND-object->BIND-object-type-id->BIND-small-molecule-id BIND-object->BIND-object-type-id->Molecular-complex-id. 5. All molecules will be identified by either an alphanumeric accession number (see 'A-acc' and 'B-acc' for interaction records and 'acc' for complex records) or by an integer identifier (see 'A-id' and 'B-id' for interaction records and 'id' for complex records) or by both; the identifiers listed for a molecule depend upon the source database. For example, molecules descibed by GenBank will have an entry in both the 'acc' and 'id' columns (corresponding to the accession number and GenInfo (GI) identifier of the sequence record). On the other hand, molecules descibed in KLOTHO, LIGAND and MMDB will have only an alphanumeric idenifier and a zero (0) will be listed in place of an integer identifier in the 'id' column. Finally, complexes described in BIND will have only an integer identifier and '-' will appear in place of an alphanumeric identifier in the 'acc' column. There may be cases when an alphanumeric GenBank accession number is expected but the letters 'NA' appear instead. These entries represent BIND records that reference a retired GenBank record (for example, see GI 6321066 in GenBank). BIND records are reviewed on an ongoing basis to replace these retired GI's with the more up-to-date GI and accession identifiers. Therefore, these 'NA' entries will disappear over time. 6. Fields named 'A-acc' , 'B-acc' or 'acc' contain alphanumeric record identifiers from the database (specified by 'A-db', 'B-db' or 'db' field) for the records describing molecule A or B (if you are looking at an interaction record) or a subunit of a complex (if you are looking at a complex record). 7. Fields named 'A-id', B-id or 'id' contain integer identifiers that point to records in some database (see corresponding column A-db , B-db or id). This record describes a molecule A or B (if you are looking at an interaction record) or a subunit of a complex (if you are looking at a complex record). 8. Fields named 'A-tax' , 'B-tax' or 'tax' contain integer taxon identifiers for the molecule described in the adjacent columns. This field is required where the type of molecule is one of DNA, RNA, protein or gene; otherwise this field may contain '0' (zero). The file "yyyymmdd.taxon.txt" contains a list of taxon identifiers and corresponding scientific names. 9. Note that the A and B molecules may not be listed in the same order in both the redundant group record and any one of the individual interaction records belonging to the group. The value of the column AB answers the question: "Are A and B listed in the same order in both the redundant group record and the individual interaction record table ("yes" or "no")." 10. PubMed Identifier of literature reference that supports this interaction. There will be at least one pmid entry for every BIND record identifier in the table. The same pmid may be listed more than once for the same BIND record identifier if a different method is listed in each case. Therefore, any combination of bid, pmid and method will not be found more than once in this table. PubMed Identifiers will not in this index if (1) they do not support the interaction and (2) if they are not found in one of the following BIND record objects: BIND-condition-set BIND-pub-set BIND-place-set BIND-loc 11. Method used in paper to provide evidence for the interaction. This field is a choice of: not-specified, alanine-scanning, affinity-chromatography, atomic-force-microscopy, autoradiography, competition-binding, cross-linking, deuterium-hydrogen-exchange, electron-microscopy, electron-spin-resonance, elisa, equilibrium-dialysis, fluorescence-anisotropy, footprinting, gel-retardation-assays, gel-filtration-chromatography, hybridization, immunoblotting, immunoprecipitation, immunostaining, interaction-adhesion-assay, light-scattering, mass-spectrometry, membrane-filtration, monoclonal-antibody-blockade, nuclear-translocation-assay, phage-display, reconstitution, resonance-energy-transfer, site-directed-mutagenesis, sucrose-gradient-sedimentation, surface-plasmon-resonance-chip, transient-coexpression, three-dimensional-structure, two-hybrid-test, allele-specific-complementation, far-western, colocalization, other For more information, see the BIND specification: BIND-experimental-system. 12. The "nrints" table is essentially the same as the "ints" but interactions between any two biomolecules are only listed once. Each molecule (A and B) are represented by the "best annotated" record for that molecule. 13. Taxon identifier and name are from NCBI's taxonomy database. (see http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Taxonomy). For example, taxon id 4932 refers to "Saccharomyces cerevisiae". 14. Every BIND record that describes a complex of molecular subunits is represented in this table once. 15. The number of differnt types of subunits contained in the complex. Individual subunit types are described in the file yyyymmdd.subunits.txt. 16. A text description of the molecular complex. 17. Each row in this table represents a subunit of the complex. The bcid (BIND complex record identifier) will appear once in this table for every different type of subunit belonging to the complex. 18. This table lists interactions that occur within complexes. In the BIND data model, complexes are described as collections of interaction records that make up the complex. These interactions may be between two known molecular objects. If this is the case then the listed bid (BIND interaction record id) will also appear in the BIND interaction index (see yyyymmdd.ints.txt). However, in many cases, the subunits of the complex are known but it is unknown how those subunits interact with one another (i.e. complexes with unknown topologies). In this case, the complex record will be composed of a series of interaction records where one of the two objects is known and the other object is of the type "not-specified". These interactions will NOT appear in the BIND interaction index (see yyyymmdd.ints.txt). However, a description of the subunit types will appear for all complexes (known and unknown topologies) in the BIND complex subunits index (yyyymmdd.complex2subunits.txt). 19. This table lists all possible interactions between all subunits in a complex where the topology of the complex is undefined (see note 18). This table is only a representation of a complex record; the interactions listed may or may not exist. The total number of all possible combinations of subunits is calculated as an arithmetical progression with the following formula: total_number = x(x+1)/2 where: 'total_number' = the total number of all possible interactions and 'x' = the total number of subuints within the complex minus 1. 20. This table lists potential interactions between all subunits of a complex and a central complex that was used to experimentally demonstrate the existence of a complex; for example, the subunit that was used as 'bait' in an immunoprecipitation experiment. This table is only a representation of a complex record; the interactions listed may or may not exist. The total number of interactions for a complex will be: total_number = x - 1 where 'total_number' = the total number of listed interactions and x is total number of subunits within the complex. 21. This table lists all PubMed Identifiers (PMIDs) belonging to a given complex record (bcid). PMID's must support the complex record to be included in this list. Additional references may exist in the interaction records that make up the complex record (see note 18); these references will be listed in the 'BIND interaction references and methods index' (yyyymmdd.refs.txt). 22. A short label for the molecule. For details see BIND-object->short-label in the BIND specification. 23. A list of short-label synonyms for the molecule, separated with "|" symbol. For details see BIND-object->other-names in the BIND specification. Note that this field may contain a list of names separated by semi-colons. This is not an error in the construction of the index file but simply reflects how data was entered into rge free-text field in the original BIND record. 24. This column is not for parsing. It is meant only to provide a human-readable description of a reference that is listed in a BIND record. Please, view this file in UTF-8 encoding. The format of the line is as following: LastName, Initials., LastName, Initials. ArticleTitle. Year. JournalTitle Volume (Issue): pp-pp "LastName" contains the surname. "Initials" contains up to two initials. Please note the these two fields can occur more then once. "ArticleTitle" contains the entire title of the journal article. "Year" contains the year on which the issue of the journal was published. "JournalTitle" contains the standard abbreviation for the title of the journal in which an article appeared. "Volume" is the volume number of the journal in which the article was published. "Issue" identifies the issue, part or supplement of the journal in which the article was published. "pp-pp" indicates the inclusive pages for the article cited. The pagination can be entirely non-digit data and redundant digits are omitted. More details may be found at http://www.nlm.nih.gov/bsd/licensee.html. 25. The best annotated GI for a given group of redundant protein sequences. For a more detailed description, please see the function description for SHoundFirstOfRedundantGroupFromID at http://www.blueprint.org/seqhound/api_help/apifunctslist.html ================================================================================ end of notes to BIND index table descriptions ================================================================================ If you use BIND, please cite: Bader GD, Betel D, Hogue CW. (2003) BIND: the Biomolecular Interaction Network Database. Nucleic Acids Res. 31(1):248-50 PMID: 12519993 The Blueprint Initiative Samuel Lunenfeld Research Institute Mt. Sinai Hospital 522 University Ave. Toronto, Ontario, Canada M5G 1W7 Phone: 416-596-8505 Fax: 416-596-8077 BIND Inquires: info@bind.ca SeqHound Inquires: seqhound@blueprint.org TraDES Inquiries: trades@blueprint.org Other: info@blueprint.org DATA POLICY--------------------------------------------------------------------- BIND data is free for both commercial and academic use. If you use BIND data, please cite: Bader GD, Betel D, Hogue CW. (2003) BIND: the Biomolecular Interaction Network Database. Nucleic Acids Res. 31(1):248-50 PMID: 12519993 This data is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. BIND DIVISIONS Get the BIND Division files when you want to assemble a complete non-redundant BIND database. The BIND Fungi division is currently the largest in BIND. Data is provided on the BIND FTP () site in a variety of formats. As the size of a database reaches critical mass, a method must be developed by which data in the database is categorized. This facilitates the production of data files of practical size for distribution. The categories developed must partition the data in a non-redundant manner. That is, the collection of all data files from divisions must contain one and only one copy of every record in the database. We have established the following divisions based on the computer readable instances of the biological information and taxonomy: * BIND Taxroot|Record where all molecules are from any organism except those within TaxID 4751 or TaxID 33208 * BIND Fungi|Record where all molecules are from TaxID 4751 * BIND Metazoa|Record where all molecules are from TaxID 33208 * RefBIND Taxroot|Reference quality record where all molecules are from any organism * PDB MMDB MMDBBIND 3DBP Taxroot|Auto-generated record where all molecules are protein, DNA or RNA and are from any organism * PDB MMDB MMDBBIND 3DSM Taxroot|Auto-generated record where at least one molecule is a small molecule and molecules are from any organism. BIND DATA SETS Get the BIND Data Set files corresponding to specific BIND entries, grouped by taxonomy, by published paper or by experiment. Use BIND Data Sets to compare high-throughput experiments or data from different organisms. Data is provided on the BIND FTP site () in a variety of formats. As requested by our users, we also provide data files containing sets of related data in order to facilitate data analysis. Users must be aware that the collection of data files from data sets contains redundant records. Note that we have not yet at the time of writing included the MMDBBIND records into BIND data sets. We have established data set categories based on the following: 1. Publication One dataset in this category contains BIND Interaction records that reference the same publication when that publication is referenced in 10 or more records. Thus, this category contains all high-throughput datasets in BIND. 2. Taxonomy One dataset in this category contains BIND Interaction records in which one or both molecules are from the specified organism. Thus, intra-organismal interactions will be found in the datasets from both organisms. For example, a virus-host interaction would be found in both the virus and host datasets. 3. Experimental system One dataset in this category contains BIND Interaction records in which the specified experimental system is identified one or more times. Thus, if one BIND record contains results from both yeast two-hybrid and co-immunoprecipitation experiments, the BIND record will be found in both the yeast two-hybrid and co-immunoprecipitation datasets. HIGH-THROUGHPUT DATA By default, BIND displays complexes with unknown interaction topology using a spoke (bait binds to prey) model, which we have shown to be three times more accurate than the (all-binding-all) matrix data model. Certain high-throughput (HTP) techniques for studying protein interactions are insufficient in their experimental approach to distinguish between direct and indirect interactions. More specifically, identifying proteins by mass spectrometry that were isolated from a cell or tissue lysate via immunoprecipitation of a protein of interest does not establish that each of the proteins identified (i.e. prey) bound directly to the protein of interest (i.e. bait). It may be that one or more proteins mediate the association of the bait and prey permitting the prey to be identified via immunoprecipitation of that bait. The connections between molecules in a complex are referred to as topology. In order to elucidate topology for a given complex, data from multiple experimental techniques, for example reciprocal immunoprecipitation experiments or yeast two-hybrid assays, are required. In the absence of further experimental results, data from HTP immunoprecipitation experiments must be interpreted using a data model, which includes assumptions. Two common data models for representing HTP immunoprecipitation results are the spoke model and the matrix model (Bader and Hogue, Nature Biotechnology 2002 20:991). The spoke model assumes that each prey protein interacts directly with the bait protein. A schematic of this topology resembles a wheel; the bait protein is the hub and the lines to each prey protein representing the interaction are spokes (Figure 1). The matrix model assumes that each protein identified in the complex interacts with each of the other proteins found in that complex. Thus, a two dimensional matrix of protein-protein interactions would have each protein listed on each of the two dimensions and every possible pair-wise interaction would be identified (Table 1). Our own comparison of results from HTP immunoprecipitation experiments with interactions previously identified in the literature demonstrates that the spoke model is 3 times more accurate than the matrix model for the representation of HTP immunoprecipitation data. Therefore, the spoke model is used in BIND to visualize HTP immunoprecipitation results. However, since the direct interaction partners are not identified, BIND does not store an Interaction record containing two known proteins rather a Complex record made up of Interaction records that each describe one bait or prey protein identified in the immunoprecipitation experiment. Specifically, data from Ho et al., 2002 and Gavin et al., 2002 are stored in BIND in this manner. BIND Complex 11509 contains the experimental results from one immunoprecipitation experiment described by Ho et al., 2002. The bait, Kss1, isolated eight other proteins as identified by immunoprecipitation and mass spectrometry. Therefore, there are nine Interaction records in BIND Complex 11509. BIND Interaction 6558 describes the bait, Kss1. This is indicated by the fact that Kss1 is Molecule A of Interaction 6558. Each of the eight prey proteins is described as Molecule B in separate BIND Interaction records (BIND ID 6559, 6560, 6561, 6562, 6563, 6564, 6565, 6566). Each of the nine proteins is recorded as interacting with Unknown, an unspecified organismal molecule. The visualization tool for viewing BIND Interactions displays this complex in the spoke model representation (Figure 1). For these two data sets, there is a difference in the number of BIND records that are counted. The statistics as shown on our webpage include the number of BIND complex records, i.e. 797 complexes for Ho et al., 2002, 454 complexes for Gavin et al., 2002. However, since the visualization and analysis of these interactions is conducted using the spoke model, the statistics for our ftp site include the number of equivalent BIND Interaction records, i.e. 3225 for Ho et al., 2002, 3618 for Gavin et al., 2002. DELETED SEQUENCE RECORDS From time to time, sequence records are removed from NCBI's GenBank database. If a BIND Interaction record references a sequence record that has been removed, the BIND record is no longer valid. For this reason, NCBI sequence record identifiers, i.e. GIs, must be monitored and potentially updated. If a GI is removed and replaced with a new record, we will automatically update the corresponding BIND record(s) to refer to the new GI. --------------------------------------------------------------------------------