******************************************************************************
RefSeq-release2.txt ftp://ftp.ncbi.nih.gov/refseq/release/release-notes/
NCBI Reference Sequence (RefSeq) Database
Release 2
October 21, 2003
Distribution Release Notes
Release Size:
2124 organisms, 7745398573 nucleotide bases, 286957682 amino acids, 1097404 records
******************************************************************************
This document describes the format and content of the flat files that
comprise releases of the NCBI Reference Sequence (RefSeq) database.
Additional information about RefSeq is available at:
1. NCBI Handbook:
http://www.ncbi.nlm.nih.gov/books/bv.fcgi?call=bv.View..ShowTOC&rid=handbook.TOC&depth=2
2. RefSeq Web Site:
http://www.ncbi.nih.gov/RefSeq/
If you have any questions or comments about RefSeq, the RefSeq release files
or this document, please contact NCBI by email at:
info@ncbi.nlm.nih.gov.
To receive announcements of future RefSeq releases and large updates please
subscribe to NCBI's refseq-announce mail list:
send email to refseq-announce-subscribe@ncbi.nlm.nih.gov
with "subscribe" in the subject line (without quotes)
OR
subscribe using web interface at:
http://www.ncbi.nlm.nih.gov/mailman/listinfo/refseq-announce
=============================================================================
TABLE OF CONTENTS
=============================================================================
1. INTRODUCTION
1.1 Release 2
1.2 Cutoff date
1.3 RefSeq Project Background
1.3.1 Sequence accessions, validation, and annotations
1.3.2 Data assembly, curation, and collaboration
1.3.3 Biologically non-redundant data set
1.3.4 RefSeq and DDBJ/EMBL/GenBank comparison
1.4 Uses and applications of the RefSeq database
2. CONTENT
2.1 Organisms included
2.2 Molecule Types included
2.3 Known Problems, Redundancies, and Inconsistencies
2.4 Last genome update for select major organisms
2.5 Release Catalog
3. ORGANIZATION OF DATA FILES
3.1 FTP Site Organization
3.2 File Names and Formats
3.3 File Sizes
3.4 Statistics
3.5 Release Catalog
3.6 Accession Format
3.7 Growth of RefSeq
4. FLAT FILE ANNOTATION
4.1 Main features of RefSeq Flat File
4.1.1 LOCUS, DEFLINE, ACCESSION, KEYWORDS, SOURCE, ORGANISM
4.1.2 REFERENCE, DIRECT SUBMISSION, COMMENT
4.1.3 FEATURE ANNOTATION (Gene, mRNA, CDS, Variation, Protein)
4.2 Tracking Identifiers
4.2.1 GeneID and LocusID
4.2.2 Transcript ID
4.2.3 Protein ID
4.2.4 Conserved Domain Database (CDD) ID
5. REFSEQ ADMINISTRATION
5.1 Citing RefSeq
5.2 RefSeq Distribution Formats
5.3 Other Methods of Accessing RefSeq Data
5.4 Request for Corrections and Comments
5.5 Credits and Acknowledgements
5.6 Disclaimer
=============================================================================
1. INTRODUCTION
=============================================================================
The NCBI Reference Sequence Project (RefSeq) is an effort to provide the
best single collection of naturally occurring biomolecules, representative
of the central dogma, for each major organism. Ideally this would include
one sequence record for each chromosome, organelle, or plasmid linked on a
residue by residue basis to the expressed transcripts, to the translated
proteins, and to each mature peptide product. Depending on the organism, we
may have some, but not all, of this information at any given time. We
pragmatically include the best view we can from available data.
1.1 Release 2
-------------
The National Center for Biotechnology Information (NCBI) at the National
Library of Medicine (NLM), National Institutes of Health (NIH) is
responsible for producing and distributing the RefSeq Sequence Database.
Records are provided through a combination of collaboration and in-house
processing including some curation by in-house staff comprised of expert
biologists.
RefSeq Release 2 is a full release of all NCBI RefSeq records.
The RefSeq project is an ongoing effort to provide a curated, non-redundant
collection of sequences. This first release includes all of the sequence data
that we have collected at this time. Although the RefSeq collection is not yet
complete, its value as a non-redundant dataset has reached a level that
justifies providing full releases.
1.2 Cutoff date
---------------
This full release, Release 2, incorporates data available as of October 21, 2003.
For more recent data, users are advised to:
. Download the RefSeq daily update files from the RefSeq FTP site
ftp://ftp.ncbi.nih.gov/refseq/daily/new/
. Use the interactive web Entrez Query systems to query based on
date
http://www.ncbi.nih.gov/Entrez/
Notice of Change: a new directory has been created (daily/new/) to provide
daily updates in the same file formats as are made available with the release.
The original file formats provided in the 'daily' directory will be retained
until January 2004. At that time, the daily updates will only be provided
in the file formats consistent with the release and the 'new' directory
will be removed.
1.3 RefSeq Project Background
-----------------------------
1.3.1 Sequence accessions, validation, and annotation
-----------------------------------------------------
Every sequence is assigned a stable accession, version, and gi and
all older versions remain available over time. RefSeq accessions
have a distinct format (see section 3.6); the underscore ("_") is the
primary distinguishing feature of a RefSeq accession.
DDBJ/EMBL/GenBank accessions never include an underscore.
Sequences are validated in several ways. For example, to confirm
that genomic sequence from the region of the mRNA feature really
does match the mRNA sequence itself, and that the annotated coding region
features really can be translated into the protein sequences they refer to.
Validation also checks for valid ASN.1 format. For genomes
included in the LocusLink database, validation ensures consistency
is maintained for descriptive information (symbols, gene and protein names)
between RefSeq and LocusLink records.
Each molecule is annotated as accurately as possible with the
correct organism name, the correct gene symbol for that organism,
and reasonable names for proteins where possible. When available,
nomenclature provided by official nomenclature groups is used.
Note that gene symbols are not required or expected to be unique
either across species or within a species.
1.3.2 Data assembly, curation, and collaboration
------------------------------------------------
We welcome collaborations with authoritative groups outside NCBI
who are willing to provide the sequences, annotations, or links
to phenotypic or organism specific resources. Where such collaborations
have not yet developed, NCBI staff have assembled the best view of
the organism that we can put together ourselves. In some cases, as with the
human genome, NCBI is an active participant in generating the
genome assembly and in providing reference sequences to represent
the annotated genome. For other genomes, we may compile the data
ourselves from DDBJ/EMBL/GenBank or other public sources. For instance,
we may simply select the "best" DDBJ/EMBL/GenBank record by automatic means,
validate the data format (and correct if needed), and add an essentially
unchanged copy to the RefSeq collection, attributed to the original
DDBJ/EMBL/GenBank record. In other cases we may provide a record that is very
similar to the DDBJ/EMBL/GenBank record, but to which experts at NCBI have added
corrected or additional annotation. This latter process can range
from minor technical repairs to a manually curated re-annotation of
the sequence, often in collaboration with experts outside NCBI.
Each record that has been curated, or that is in the pool for
future curation, is labeled with the level of curation it has received.
Curation status information is provided primarily for transcript and
protein records. Curation is carried out on the whole genome level
for some smaller genomes such as viral, organelle, and some microbial
genomes.
Curation status codes are defined in the section 3.2 below.
1.3.3 Biologically non-redundant data set
-----------------------------------------
RefSeq provides a biologically non-redundant set of sequences for database
searching and gene characterization. It has the advantage of providing an
objective and experimentally verifiable definition of "non-redundant" in
supplying one example of each natural biomolecule per organism. The small
amount of sequence redundancy introduced from close paralogs, alternate splicing
products, and genome assembly intermediates is compensated for by the
clarity of the model. RefSeq provides the substrate for a variety of
conclusions about non-redundancy based on clustering identical sequences,
or families of related sequences, without confounding the database itself
with these more subjective assessments.
1.3.4 RefSeq and DDBJ/EMBL/GenBank comparison
---------------------------------------------
RefSeq is unique in providing a large curated database across many
organisms, which precisely and explicitly links genetic (chromosome),
expression (mRNA), and functional (protein) sequence data into an
integrated whole.
DDBJ/EMBL/GenBank also integrates DNA and protein information, and RefSeq is
substantially based on sequence records contributed to DDBJ/EMBL/GenBank.
However, RefSeq is similar to a review article in that it represents
a synthesis and summary of information by a particular group (NCBI or
other RefSeq contributors) that is based on the primary data gathered
by many others and made part of the scientific record. Also, like a
review article, it has the advantage of organizing a large body of
diverse data into a single consistent framework with a uniform set of
conventions and standards.
Note that while based on DDBJ/EMBL/GenBank, RefSeq is distinct from
DDBJ/EMBL/GenBank. DDBJ/EMBL/GenBank represents the sequence and annotations
supplied by the original authors and is never changed by NCBI or RefSeq staff.
DDBJ/EMBL/GenBank remains the primary sequence archive while RefSeq is a
summary and synthesis based on that essential primary data.
1.4 Uses and applications of the RefSeq database
------------------------------------------------
A stable, consistent, comprehensive, non-redundant database of genomes
and their products provides a valuable sequence resource for similarity
searching, gene identification, protein classification, comparative
genomics, and selection of probes for gene expression. It also acts as
molecular "white pages" by providing a single, uniform point of access
for searching at the sequence level, and by connecting the results with
a diversity of organism-specific databases or resources unique to that
organism or field.
=============================================================================
2. CONTENT
=============================================================================
2.1 Organisms included
----------------------
This release includes records representing 2124 distinct taxonomic categories,
as measured by counting the number of distinct tax_ids included in the release.
Tax_ids are provided, for all species having any amount of sequence data, by the
NCBI Taxonomy group.
The release includes species ranging from viral to microbial to eukaryotic and
includes organisms for which complete and incomplete genomic sequence data is
available.
The release does not include all species for which some sequence data is
available in DDBJ/EMBL/GenBank. The decision to generate RefSeq data for a
species depends in part on the amount of sequence data available.
Additional species will be represented in the RefSeq collection as
more sequence data becomes available.
2.2 Molecule Types Included
---------------------------
The RefSeq release includes genomic, transcript, and protein sequence data;
however, these molecule types are not provided for all organisms and the
sequences provided may not be complete or comprehensive for some species.
Transcript RefSeq records may represent protein-coding transcripts or
non-coding RNA products; these records are currently only provided for
eukaryotic species.
Genomic RefSeq records are provided when a sufficient quantity of genomic
sequence data is available in DDBJ/EMBL/GenBank. Transcript and protein
records may be provided for a species before genomic sequence data is available,
as is the case with Danio rerio (zebrafish).
2.3 Known Problems, Redundancies, and Inconsistencies
------------------------------------------------------
The RefSeq collection is an ongoing project that is expected to grow
in scope and content over time. Thus it is important to recognize that
it is not complete in that some genomes are not yet completely sequenced,
some incompletely sequenced genomes may not be included, or some gene
products may not yet be represented. RefSeq records may be added, removed,
or updated in future releases as new information becomes available and as
a result of curation.
Genomes with pending updates:
Sequence updates are planned in the near future for the following species.
These updates may revise the genomic, transcript, and protein data provided
in the RefSeq collection. The updated RefSeqs will be available in a future
release.
Mus musculus genome assembly and associated models
Known Data inconsistencies:
[1] RefSeq status codes are not consistently provided for some species.
The goal is to consistently provide a status code for all RefSeq
records. The release catalog indicates "UNKNOWN" if a status code
was expected but not detected and "na" if a status code is not
expected based on the original project plan for provision of this type
of information. Status codes will be more consistently applied to all
records in the future.
[2] The genomic, transcript, and protein collection is known to be
incomplete for many species. This is particularly true for those
genomes for which a complete genome assembly is not yet available,
such as Danio rerio (zebrafish), Bos taurus (cow), and Leishmania
major. As additional sequence data becomes available, the RefSeq
representation for these, and other, organisms will increase.
[3] Although the goal is to provide a non-redundant collection, some
redundancy is included in this release as follows:
Known Duplication:
Two versions of NM_172381 and NP_759013 are included in the release files
due to a processing problem.
Redundant Protein records:
Alternate Splicing When additional transcripts are provided
to represent alternate splicing products,
and the alternate splice site occurs in
the UTR, then the protein is redundantly
provided.
Paralogs The goal is to provide a RefSeq record
for each naturally occuring molecule.
Therefore, records are provided for all
genes identified including those produced
by more recent gene duplication events in
which the genes are nearly identical.
Redundant Genomic records:
Intermediate records For some species, intermediate genomic
records are provided to support the
assembly and/or annotation of the genome.
For example, for human, a chromosome may
be represented by a chromosome RefSeq
record with a NC_ accession prefix.
The chromosome record may consist of
many contigs, each represented as a
separate record with a NT_ accession
prefix. In addition, some curated gene
region records, with NG_ accession
prefix, may also be provided to support
annotation of complex regions.
Alternate assemblies Genomic records are provided to represent
alternate assemblies of genomic sequence
derived from different populations. These
records will have varying levels of
redundancy and represent polymorphic and
haplotype differences in terms of the
sequence and annotation.
For example, alternate assemblies are
provided for different mouse strains and
for regions of the human major
histocompatibility complex (MHC). The MHC
is a highly variable region of chromosome
6 which exhibits variation at the level
of both sequence polymorphism and gene
content. The alternate assemblies make it
possible to represent this alternate gene
content.
2.4 Notes on select major organisms
-----------------------------------
Anopheles gambiae Genomic sequence data is available as
whole genome shotgun (WGS).
Arabidopsis thaliana An update of the annotated genome was provided
by TIGR in July, 2003. The RefSeq release includes
chromosomes, transcripts, and proteins.
Caenorhabditis elegans The RefSeq release includes an annotation update
that was released on October 16, 2003 the genome
version available on March 7, 2003. The release
includes chromosome, transcript, and protein records.
Drosophila melanogaster Release 3.1 of the assembled, annotated genome
was provided by FlyBase in March 2003.
Homo sapiens NCBI provides the human genome assembly in close
collaboration with the sequencing centers.
RefSeq release 2 includes human genome build 34,
which is based on data available on July 30 2003.
Release 2 includes RefSeq chromosomes, contigs,
known transcripts and proteins (as defined by
having a Locus ID), and derived model transcripts
and proteins predicted by the Genome Annotation
pipeline. See:
http://www.ncbi.nlm.nih.gov/genome/guide/build.html
Mus musculus NCBI provides the mouse genome assembly in close
collaboration with the sequencing centers.
This RefSeq release includes mouse genome build 30
which is based on data available in January, 2003.
Release 2 includes RefSeq contigs, known
transcripts and proteins, and derived model
transcripts and proteins predicted by the Genome
Annotation pipeline. A mouse genome update is
imminent at the time of this release processing
and will be included in the next release.
Neurospora crassa This release adds representation for neurospora. The
annotated genome data was supplied by the Whitehead
Institute. RefSeqs were released on July 2, 2003
and include WGS genomic contigs, predicted transcripts,
predicted proteins. The RefSeq data does not represent
the subset of small WGS contigs that were not
mapped to a chromosome position or do not include
annotation.
Oryza sativa This release adds representation for rice. The
genome is being sequenced by the International
Rice Genome Sequencing Project. RefSeqs are provided
by NCBI processing to generate the annotated
genomic contigs; annotation is propagated from
the submitted BAC clones. The rice RefSeq set does not
represent annotation from BAC clones that didn't fall into
supercontigs, even though they are mapped to a chromosome
position. RefSeqs were first released on October 6, 2003 and
include genomic contigs, transcripts, and proteins.
Rattus norvegicus NCBI uses the rat whole genome shotgut (WGS)
genome assembly provided by Baylor sequencing
center. RefSeq release 2 includes rat genome
build 2 which is based on the RGSC v3.1 assembly,
provided by the Rat Genome Sequencing Consortium
(RGSC). Release 2 includes RefSeq contigs, known
transcripts and proteins, and derived model transcripts
and proteins predicted by the Genome Annotation
pipeline.
Saccharomyces cerevisiae Provided by Sacchraomyces Genome Database (SGD)
and last updated October 14, 2003. The RefSeq release
includes chromosome and protein records.
Microbial This RefSeq release includes 140 complete microbial
genomes. Microbial genomes are annotated by
a collaborative automatic computation method,
followed by curation by NCBI staff.
Fifteen microbial genomes have become available in
GenBank since the last release:
Prochlorococcus marinus subsp. marinus str. CCMP1375
Chlamydophila pneumoniae TW-183
Candidatus Blochmannia floridanus
Bordetella pertussis
Bordetella parapertussis
Bordetella bronchiseptica
Prochlorococcus marinus subsp. pastoris str. CCMP1378
Prochlorococcus marinus str. MIT 9313
Synechococcus sp. WH 8102
Chromobacterium violaceum ATCC 12472
Porphyromonas gingivalis W83
Wolinella succinogenes
Gloeobacter violaceus
Photorhabdus luminescens subsp. laumondii TTO1
Vibrio vulnificus YJ016
Seventeen microbial genomes have been curated:
Aeropyrum pernix
Archaeoglobus fulgidus
Buchnera sp. APS
Buchnera aphidicola Sg
Corynebacterium glutamicum
Escherichia coli K-12
Haemophilus influenzae
Lactococcus lactis subsp. lactis
Mycoplasma genitalium
Mycoplasma pneumoniae
Oceanobacillus iheyensis
Shewanella oneidensis
Pyrococcus abyssi
Pyrococcus furiosus
Pyrococcus horikoshii
Thermoplasma volcanium
Vibrio vulnificus CMCP6
Viruses This RefSeq release includes over 1231 distinct viral
records which have been curated via an extensive
collaboration between the international virologist
community and NCBI staff virologists. A panel of
viral genomes advisors has been established.
For more information please see:
RefSeq Collaborations: http://www.ncbi.nih.gov/RefSeq/collaborators.html
Viral Genome Advisors: http://www.ncbi.nih.gov/PMGifs/Genomes/viradvisors.html
Microbial Contributors: http://www.ncbi.nih.gov/RefSeq/microbialcontrib.html
2.5 Release Catalog
-------------------
The Release Catalog documents the full contents of the RefSeq Release.
The catalog can be used to identify data of interest. See the format
description in section 3.5 for additional information.
The release catalog is available at:
ftp://ftp.ncbi.nih.gov/refseq/release/release-catalog/RefSeq-release2.catalog
=============================================================================
3. ORGANIZATION OF DATA FILES
=============================================================================
3.1 FTP Site Organization
-------------------------
RefSeq releases are available on the NCBI FTP site at:
ftp://ftp.ncbi.nih.gov/refseq/release/
The RefSeq collection is provided in a redundant fashion to best meet the needs
of those who want the full collection as well as those who want a specific
sub-set of the collection. Therefore the collection is provided as:
1) the complete collection, and
2) sections as defined by major taxonomic or other logical groupings.
A subdirectory exists for each sub-section as follows:
fungi
invertebrate
microbial
mitochondrion
plant
plasmid
plastid
protozoa
vertebrate_mammalian
vertebrate_other
viral
In addition, the complete collection is available without these
sub-groupings in the subdirectory:
complete
Note that this directory structure intentionally provides the release
data in a redundant fashion. We gave considerable thought to how to
package the release to meet the needs of different user groups.
For instance, some groups may be interested in retrieving the complete
protein set, while other groups may be interested in retrieving data
for a more limited number of organisms. We decided to provide
logical groupings based on general taxonomic node (viral, mammalian etc)
as well as logical molecule type compartmentalization (e.g., plastid).
Thus, all records are provided at least twice, once in the "complete"
directory, and a second time in one of the other directories.
Some sequences may be provided three times when it is logical to
include the record in more than one additional directory. For example,
a sequence may be provided in the "complete", "mitochondrion", and
"vertebrate_mammalian" directories.
We are interested in hearing if you find this structure useful or if
you would like information grouped in a different manner.
Send suggestions or comments to the NCBI Help Desk at:
info@ncbi.nlm.nih.gov
3.2 File Names and Formats
--------------------------
File names are informative, and indicate the content, molecule type,
and file format of each RefSeq release data file. Most filenames
utilize this structure:
directoryfilenumber.molecule.format.gz
1 2 3 4
File Name Key:
1. directory directory level the file is provided in
(e.g.,complete, viral etc)
2. file number: large data sets are provided as incrementally
numbered files
3. molecule type of molecule (genomic, rna, or protein);
not relevant for ASN.1 format files provided
in the "complete" sub-directory
4. format the data format provided in the file; see below
For example:
complete1.genomic.bna.gz
vertebrate_mammalian2.protein.gpff.gz
RefSeq Whole Genome Shotgun (WGS) data are provided in files provided
per WGS project. Their filenames use a slightly different structure:
directoryWGSproject.molecule.format.gz
For example:
completeNZ_AAAU.bna.gz
microbialNZ_AAAV.genomic.fna.gz
All RefSeq release files have been compressed with the gzip utility;
therefore, an invariant ".gz" suffix is present for all release files.
The data that comprises a RefSeq release are available in several
file formats, as indicated by the format component in the file name:
bna binary ASN.1 format; includes nucleotide and protein
gbff GenBank flat file format; nucleotide records
gpff GenPept flat file format; protein records
fna FASTA format; nucleotide records
faa FASTA format; protein records
The comprehensive full release is deposited in the "complete"
directory and is available in all file types.
Binary ASN.1 format is only provided in the complete directory. The remaining
directory include all of the remaining file types.
The DDBJ/EMBL/GenBank and GenPept flat file format provided in this release
matches that seen when accessing the records using the NCBI web site.
Notably, some RefSeq record are in the CON division and do not instantiate
the sequence on the flat file display, instead a 'join' statement is provided
to indicate the assembly instructions. The FASTA files do include the
assembled sequences for these CON division RefSeq records.
For example, see NC_000022.
Suggestions regarding the structure of the RefSeq release product
and the available formats may be sent to the NCBI Help Desk:
info@ncbi.nlm.nih.gov
3.3 File Sizes
--------------
RefSeq release files are provided in a range of sizes. Most are
limited to several hundred megabytes. However, some of the genomic
FASTA files can exceed 2Gb.
Files are compressed to reduce file size and facilitate FTP retrieval.
The total size of release 2 (includes all directories) is as follows:
Extension Size (GB) Type
-----------------------------------------------------------
bna 4.45 ASN.1
gbff 7.89 GenBank flat file
gpff 4.95 GenPept flat file
fna 34.61 FASTA, nucleotide
faa 0.69 FASTA, protein
Note: for release 2, the compete directory provides all file types. The ASN.1
format is only available in the complete directory; the file sizes reported for
the remaining file formats represents the redundant total found in the complete
plus other directories.
3.4 Statistics
---------------
RefSeq release 2 includes sequences from 2124 different organisms.
The number of species represented in each Release sub-directory,
determined by counting distinct tax IDs, is as follows:
complete 2124
fungi 34
invertebrate 81
microbial 378
mitochondrion 437
plant 33
plasmid 37
plastid 34
protozoa 40
vertebrate_mammalian 82
vertebrate_other 208
viral 1231
Total Number of Accessions, Length (number of nucleotides or amino
acids, per type of molecule:
Type Accessions Length
------------------------------------------
Genomic: 64805 7399000384
RNA: 201312 346398189
Protein: 831287 286957682
Complete RefSeq release statistics for each directory are provided
in a separate document. Please see:
ftp://ftp.ncbi.nih.gov/refseq/release/release-statistics/
file: RefSeq-release2.10212003.stats.txt
3.5 Release Catalog Format
--------------------------
The full non-redundant contents of the release are documented in the
release catalog.
The catalog includes the following columns:
1. tax_id
2. species name
3. RefSeq accession.version
4. gi
5. FTP directories data is provided in
6. RefSeq status code
7. sequence length
Note: the molecule type for each catalog entry can be inferred from
the accession prefix (see below).
RefSeq Status Codes are documented on the RefSeq web site. The catalog for
release 2 includes the following terms:
na Not Applicable;
status codes are not provided for some genomic records
UNKNOWN The status code has not yet been applied
REVIEWED The RefSeq record has been the reviewed by NCBI
staff or by a collaborator. Some RefSeq records
may incorporate expanded sequence and annotation
information including additional publications
and features.
VALIDATED The RefSeq record has undergone an initial review
to provide the preferred sequence standard. The
record has not yet been subject to final review
at which time additional functional information
may be provided.
PROVISIONAL The RefSeq record has not yet been subject to
individual review and is thought to be well
supported and to represent a valid transcript
and protein.
PREDICTED The RefSeq transcript may represent an ab initio
prediction or may be partially supported by other
transcript data; the protein is predicted.
INFERRED The RefSeq record is inferred by genome sequence
analysis.
MODEL RefSeq records provided via automated processing
and are not subject to individual review or revision
between builds.
3.6 RefSeq Accession Format
---------------------------
RefSeq accessions are formatted as a two letter prefix, followed by
an underscore, followed by six digits or 4 letters plus eight digits.
For example, NM_020236 and NZ_AABC02000001.
The underscore ("_") is the primary distinguishing feature of a RefSeq
accession; DDBJ/EMBL/GenBank accessions never include an underscore.
The RefSeq accession prefix indicates the molecule type.
Molecule Type Accession Prefix
----------------------------------------------
protein NP_; XP_; ZP_
rna NM_; NR_; XM_; XR_
genomic NC_; NG_; NT_; NW_; NZ_
Additional information is available on the RefSeq Web site:
http://www.ncbi.nih.gov/RefSeq/key.html#accessions
NOTICE OF CHANGE:
NP_ accession space will need to be expanded in the near future,
the new format will be NP_12345678. Existing accessions will remain
unchanged. That is, existing accessions, such as NP_013474, will not
be modified to 8 digits (NP_013474 and NP_00013474 will be distinct
accessions identifying different protein records)
As other accession series need to be expanded, they will also be
expanded by adding 2 digits with existing accessions remaining stable.
3.7 Growth of RefSeq
--------------------
Release Date Species Nucleotides Amino Acids Records
1 Jun 30, 2003 2005 4672871949 263588685 1061675
2 Oct 21, 2003 2124 7745398573 286957682 1097404
=============================================================================
4. FLAT FILE ANNOTATION
=============================================================================
4.1 Main features of RefSeq Flat File
-------------------------------------
Also see the RefSeq web site and the NCBI Handbook, RefSeq chapter.
http://www.ncbi.nih.gov/RefSeq/
http://www.ncbi.nlm.nih.gov/books/bv.fcgi?call=bv.View..ShowTOC
&rid=handbook.TOC&depth=2
4.1.1 LOCUS, DEFLINE, ACCESSION, KEYWORDS, SOURCE, ORGANISM
--------------------------------------------------------------------
The beginning of each RefSeq records provides information about the accession,
length, molecule type, division, and last update date. This is followed by the
descriptive DEFINITION line, then by the Accession, version,and GI data,
followed by detailed information about the organism and taxomonic lineage.
//
LOCUS NC_004916 384518 bp DNA linear INV 26-JUN-2003
DEFINITION Leishmania major chromosome 3, complete sequence.
ACCESSION NC_004916
VERSION NC_004916.1 GI:32189699
KEYWORDS .
SOURCE Leishmania major
ORGANISM Leishmania major
Eukaryota; Euglenozoa; Kinetoplastida; Trypanosomatidae;
Leishmania.
//
Note: Both the GI and VERSION number increment when a sequence is updated,
while the ACCESSION remains the same. The GI and "ACCESSION.VERSION"
identifiers provide the finest resolution reference to a sequence.
4.1.2 REFERENCE, DIRECT SUBMISSION, COMMENT
-------------------------------------------
REFERENCE:
While the majority of RefSeq records do include REFERENCE
data, this data is not required and some records do not include any
citations. Publications are propagated from the GenBank record(s) from which
the RefSeq is derived, provided by collaborating groups and NCBI staff
during the curation process, and provided by the National Library of
Medicine (NLM) PubMed MeSH indexing staff as they add new articles to PubMed.
Functionally relevant citations are added by individual scientists using
the LocusLink GeneRIF submission form, and a significant volume of citation
connections are supplied by the NLM MeSH indexing staff for human,
mouse, rat, zebrafish,and cow. This functionality is expected to increase
in the future to treat all organisms represented in the RefSeq collection.
Citations supplied by the MeSH indexers and individual scientists can be
identified by the presence of a REMARK beginning with the text string "GeneRIF".
This represents a significant method to keep sequence connections to the
literature up-to-date; GeneRIFs add considerable value to the RefSeq
collection.
For more information on GeneRIFs please see:
http://www.ncbi.nlm.nih.gov/LocusLink/GeneRIFhelp.html
For example, several GeneRIFs have been added to NM_000173.1 including:
//
REFERENCE 13 (bases 1 to 2480)
AUTHORS Poujol,C., Ware,J., Nieswandt,B., Nurden,A.T. and Nurden,P.
TITLE Absence of GPIbalpha is responsible for aberrant membrane
development during megakaryocyte maturation: ultrastructural study
using a transgenic model
JOURNAL Exp. Hematol. 30 (4), 352-360 (2002)
MEDLINE 21935100
PUBMED 11937271
REMARK GeneRIF: Absence of GPIbalpha is responsible for aberrant membrane
development during megakaryocyte maturation; leads to abnormal
partitioning of the membrane systems and abnormal proplatelet
production.
//
DIRECT SUBMISSION:
A Direct Submission field is provided on some RefSeq records but not all. It
is propagated from the underlying GenBank record from which the RefSeq is
derived or provided on submissions from collaborating groups. Transcript
and protein RefSeqs for human, mouse, rat, zebrafish, and cow do not provide
this field as records often include additional data and are not necessarily
direct copies of the GenBank submission.
COMMENT:
A COMMENT identifying the RefSeq Status is provided for the majority of the
RefSeq records. This comment may include information about the RefSeq status,
collaborating groups, and the GenBank records(s) from which the RefSeq is
derived. The RefSeq COMMENT is not provided comprehensively in this release.
We are working to supply this COMMENT more comprehensively in the future.
Additional COMMENTS are provided for some records to provide information
about the sequence function, notes about the aspects of curation, or
comments describing transcript variants.
A COMMENT is always provided if the GI has changed.
For example (from NM_133490):
//
COMMENT REVIEWED REFSEQ: This record has been curated by NCBI staff. The
reference sequence was derived from BC008969.1.
On Dec 31, 2002 this sequence version replaced gi:19424123.
Summary: Voltage-gated potassium (Kv) channels represent the most
complex class of voltage-gated ion channels from both functional
and structural standpoints. Their diverse functions include
regulating neurotransmitter release, heart rate, insulin secretion,
neuronal excitability, epithelial electrolyte transport, smooth
muscle contraction, and cell volume. This gene encodes a member of
the potassium channel, voltage-gated, subfamily G. This member
functions as a modulatory subunit. The gene has strong expression
in brain. Alternative splicing results in two transcript variants
encoding distinct isoforms.
Transcript Variant: This variant (2) has an alternate 3' sequence,
as compared to variant 1. It encodes isoform 2 that is shorter and
has a distinct C-terminus as compared to isoform 1.
//
4.1.3 NUCLEOTIDE FEATURE ANNOTATION
-----------------------------------
Gene, mRNA, CDS:
Every effort is made to consistently provide the Gene and coding sequence (CDS)
feature (when relevant). If a RefSeq is based on a GenBank record that is only
annotated with the CDS, then a Gene feature is created. mRNA features are
provided for most eukaryotic records; this is not yet comprehensively provided
and will improve in future releases.
Gene Names:
Gene symbols and names are provided by external official
nomenclature groups for some organisms. If official nomenclature is
not available we may use a systemic name provided by the data submittor
or apply a more functional name during curation. When official nomenclature
is available we may provide additional alternate names for some organisms.
Variation:
Variation is computed by the dbSNP database staff and added via post-processing
to RefSeq records.
Miscellaneous:
For some records, additional annotation may be provided when identified by the
curation staff or provided by a collaborating group. For example, the location
of polyA signal and sites may be included.
4.1.4 PROTEIN FEATURE ANNOTATION
--------------------------------
Protein Names:
Protein names may be provided by a collaborating group, may be based on the
Gene Name, or for some records, the curation process may identify the
preferred protein name based on that associated with a specific EC number
or based on the literature.
Protein Products:
Signal peptide and mature peptide annotation is provided by propagation from
the GenBank submission that the RefSeq is based on, when provided by a
collaborating group, or when determined by the curation process.
Domains:
Domains are computed by alignment to the NCBI Conserved Domain Database
database for human, mouse, rat, zebrafish, nematode, and cow. The best
hits are annotated on the RefSeq. For some records, additional functionally
significant regions of the protein may be annotated by the curation staff.
Domain annotation is not provided comprehensively at this time.
4.2 Tracking Identifiers
------------------------
Several identifiers are provided on RefSeq records that can be used to track
relationships between annotated features, relationships between RefSeq records,
and changes to RefSeq records over time.
The LocusID identifies the related Gene, mRNA, and CDS features. Transcript IDs
(RefSeq accessions) provide an explicit connection between a transcript feature
annotated on a genomic RefSeq record, and the RefSeq transcript record itself.
Likewise, the Protein ID (RefSeq accessions) provides the association between
the annotated CDS feature on a genomic or transcript RefSeq record, and the
protein record itself.
Changes to a RefSeq sequence over time can be identified by changes to the GI
and version number.
4.2.1 GeneID and LocusID
------------------------
Release 2 includes a new identifier, the GeneID. The GeneID is being added
to RefSeq records to support development of a new Entrez database, Entrez
Gene. Entrez Gene will provide gene-oriented information for the entire
RefSeq collection. It represents a significant expansion of the LocusLink
database concept. Once Entrez Gene is publicly available, then the dbxref
that is already provided on the RefSeq records will be hotlinked, in the
web Entrez GenBank and GenPept record displays, to this new resource.
The GeneID was initially set to be equivalent to the LocusID; these IDs
will not be synchronized in the future as new identifiers are added.
A distinct tracking ID (the LocusID) is available for human, mouse, rat,
zebrafish, and cow records. The LocusID is provided as a database cross-
reference qualifier (dbxref) on the gene, mRNA, and CDS features. The LocusID
can be used to identify the set of related features; this is especially useful
when multiple products are provided to represent alternate splicing events.
For example:
//
gene 19683..104490
/gene="DLEC1"
/db_xref="GeneID:9940" <<<--- GeneID
/db_xref="LocusID:9940" <<<--- LocusID
/db_xref="MIM:604050"
//
When viewing RefSeq records via the internet, the LocusID is hot-linked to the
LocusLink Gene Report page where additional descriptive information may be
available about the gene. In the future, the LocusID will be renamed "GeneID"
and will be provided for all species included in the RefSeq collection.
4.2.2 Transcript ID
-------------------
The transcript_id qualifier found on a mRNA or other RNA feature annotation
provides an explicit correspondance between a feature annotation on a genomic
record and the RefSeq transcript record.
For example:
NT_011523.9 Homo sapiens chromosome 22 genomic contig.
//
mRNA complement(231444..239103)
/gene="PKDREJ"
/product="polycystic kidney disease (polycystin) and REJ
(sperm receptor for egg jelly homolog, sea urchin)-like"
/note="Derived by automated computational analysis using
gene prediction method: BestRefseq,BLAST. Supporting
evidence includes similarity to: 3 mRNAs"
/transcript_id="NM_006071.1 <<<--- linked RefSeq transcript
/db_xref="GI:5174632"
/db_xref="GeneID:10343"
/db_xref="LocusID:10343"
/db_xref="MIM:604670"
//
4.2.3 Protein ID
----------------
The protein_id qualifier found on a coding region (CDS) feature provides an
explicit correspondance between feature annotation on a genomic or transcript
RefSeq record and the RefSeq transcript record.
For example:
NC_001144.2 Saccharomyces cerevisiae chromosome XII, complete chromosome
sequence.
//
CDS complement(16639..17613)
/gene="MHT1"
/locus_tag="YLL062C"
/note="Mht1p;
go_component: cellular_component unknown [goid 8372]
[evidence ND];
go_function: homocysteine S-methyltransferase activity
[goid 8898] [evidence IDA] [pmid 11013242];
go_process: sulfur amino acid metabolism [goid 96]
[evidence IMP] [pmid 11013242]"
/codon_start=1
/evidence=experimental
/product="S-Methylmethionine Homocysteine
methylTransferase"
/protein_id="NP_013038.1" <<<--- linked RefSeq protein
/db_xref="GI:6322966"
/db_xref="SGD:S0003985"
/db_xref="GeneID:850664"
/translation="MKRIPIKELIVEHPGKVLILDGGQGTELENRGININSPVWSAAP
FTSESFWEPSSQERKVVEEMYRDFMIAGANILMTITYQANFQSISENTSIKTLAAYKR
FLDKIVSFTREFIGEERYLIGSIGPWAAHVSCEYTGDYGPHPENIDYYGFFKPQLENF
NQNRDIDLIGFETIPNFHELKAILSWDEDIISKPFYIGLSVDDNSLLRDGTTLEEISV
HIKGLGNKINKNLLLMGVNCVSFNQSALILKMLHEHLPGMPLLVYPNSGEIYNPKEKT
WHRPTNKLDDWETTVKKFVDNGARIIGGCCRTSPKDIAEIASAVDKYS"
//
4.2.4 Conserved Domain Database (CDD) ID
----------------------------------------
The CDD identifier found on protein records, and mapped to associated
nucleotide records as a misc_feat,identifies protein domains that are
found on the record. CDD annotation is applied computationally. Initially
this annotation was provided for a subset of RefSeq; it will be applied
to the entire collection in the near future.
For example:
NP_000550.2 A-gamma globin
//
Region 5..147
/region_name="Globin"
/note="globin"
/db_xref="CDD:pfam00042" <<<--- conserved domain database
//
=============================================================================
5. REFSEQ ADMINISTRATION
=============================================================================
The National Center for Biotechnology Information (NCBI), National Library
of Medicine, National Institutes of Health, is responsible for the production
and distribution of the NIH RefSeq Sequence Database. NCBI distributes
RefSeq sequence data by anonymous FTP. For more information, you may contact
NCBI by email at info@ncbi.nlm.nih.gov or by phone at 301-496-2475.
5.1 Citing RefSeq
-----------------
When citing data in RefSeq, it is appropriate to to give the sequence name,
and primary accession and version number (or GI). Note, the most accurate
citation of the sequence is provided by including the combined accession plus
version number or the GI number.
It is also appropriate to list a reference for the RefSeq project. The
following on-line publication provides the most complete description and
should be cited when possible:
The NCBI handbook [Internet]. Bethesda (MD): National Library of
Medicine (US), National Center for Biotechnology Information; 2002
Oct. Chapter 17, The Reference Sequence (RefSeq) Project.
Available from http://www.ncbi.nih.gov/entrez/query.fcgi?db=Books
If on-line citations are not accepted by a journal, please use the following
citation:
NCBI Reference Sequence Project: update and current status
Pruitt KD, Tatusova T, Maglott DR
Nucleic Acids Res 2003 Jan 1;31(1):34-37
5.2 RefSeq Distribution Formats
-------------------------------
Complete flat file releases of the RefSeq database are available via
NCBI's anonymous ftp server:
ftp://ftp.ncbi.nih.gov/refseq/release/
Each release is cumulative, incorporating previous data plus new data.
Records that have been suppressed are not included in the release.
Incremental updates that become available between RefSeq releases
are available at:
ftp://ftp.ncbi.nih.gov/refseq/daily/new
ftp://ftp.ncbi.nih.gov/refseq/cumulative
Please refer to the README for additional information:
ftp://ftp.ncbi.nih.gov/refseq/README
ftp://ftp.ncbi.nih.gov/refseq/CHANGE_NOTICE
5.3 Other Methods of Accessing RefSeq Data
------------------------------------------
Entrez is a molecular biology database system that presents an integrated
view of DNA and protein sequence data, structure data, genome data,
publications, and other data fields. The Entrez query and retrieval
system is produced by the National Center for Biotechnology Information
(NCBI) and is available only via the internet.
Entrez is accessed at:
http://www.ncbi.nih.gov/Entrez/
RefSeq entries are indexed for retrieval in the Entrez system. The web-based
filter restrictions can be used to restrict your query to RefSeq data or to
specific subsets of the RefSeq database.
Additional specific property restrictions are provided to support querying
for RefSeq records with specific STATUS codes. Queries are defined on the
RefSeq web site at:
http://www.ncbi.nih.gov/RefSeq/
5.4 Request for Corrections and Comments
----------------------------------------
We welcome your suggestions to improve the RefSeq collection; we invite
groups interested in contributing toward the collection and curation
of the RefSeq database to improve the representation of single genes,
gene families, or complete genomes to contact us.
Please refer to RefSeq accession and version numbers (or GI) and the RefSeq
Release number (release 2) to which your comments apply; it is useful if you
indicate the source of data that you found to be problematic (e.g., data on
the FTP site, data retrieved on the web site), the entry DEFLINE, and the
specific annotation field for which you are suggesting a change.
Suggestions and corrections can be sent to:
info@ncbi.nlm.nih.gov
5.5 Credits and Acknowledgements
--------------------------------
This RefSeq release would not be possible without the support of numerous
collaborators and the primary sequence data that is submitted by thousands
of laboratories and available in GenBank.
The RefSeq project is ambitious in scope and we actively welcome opportunities
to work with other groups to provide this collection. We value all of our
collaborators; they contribute information with a large range in scope and
volume such as completely annotated genomes, advice to improve the sequence
or annotation of individual RefSeq records, information about official
nomenclature, and information about function.
In addition to the significant information collected by collaboration,
numerous NCBI staff are involved in infrastructure support, programmatic
support, and curation. RefSeq is supported by 3 primary work groups that
are associated with LocusLink, Entrez Genomes, and the Genome Annotation
Pipeline.
See the RefSeq web site for a list of collaborating groups and in-house
staff.
5.6 Disclaimer
--------------
The United States Government makes no representations or warranties
regarding the content or accuracy of the information. The United States
Government also makes no representations or warranties of merchantability
or fitness for a particular purpose or that the use of the sequences will
not infringe any patent, copyright, trademark, or other rights. The
United States Government accepts no responsibility for any consequence
of the receipt or use of the information.
For additional information about RefSeq releases, please contact
NCBI by e-mail at info@ncbi.nlm.nih.gov or by phone at (301) 496-2475.