Biological Sequence Databases

Biological Sequence Databases

Andreas D. Baxevanis

Introduction

Over the past several decades, there has been a feverish push to understand, at the most

elementary of levels, what constitutes the basic “book of life.” Biologists (and scientists in gen-

eral) are driven to understand how the millions or billions of bases in an organism’s genome

contain all of the information needed for the cell to conduct the myriad metabolic processes

necessary for the organism’s survival – information that is propagated from generation to

generation. To have a basic understanding of how the collection of individual nucleotide

bases drives the engine of life, large amounts of sequence data must be collected and stored

in a way that these data can be searched and analyzed easily. To this end, much effort has

gone into the design and maintenance of biological sequence databases. These databases have

had a significant impact on the advancement of our understanding of biology not just from

a computational standpoint but also through their integrated use alongside studies being

performed at the bench.

The history of sequence databases began in the early 1960s, when Margaret Dayhoff and

colleagues (1965) at the National Biomedical Research Foundation (NBRF) collected all of the

protein sequences known at that time – all 65 of them – and published them in a book called

the Atlas of Protein Sequence and Structure. It is important to remember that, at this point in the

history of biology, the focus was on sequencing proteins through traditional techniques such

as the Edman degradation rather than on sequencing DNA, hence the overall small number

of available sequences. By the late 1970s, when a significant number of nucleotide sequences

became available, those were also included in later editions of the Atlas. As this collection

evolved, it included text-based descriptions to accompany the protein sequences, as well as

information regarding the evolution of many protein families. This work, in essence, was the

first annotated sequence database, even though it was in printed form. Over time, the amount

of data contained in the Atlas became unwieldy and the need for it to be available in electronic

form became obvious. From the early 1970s to the late 1980s, the contents of the Atlas were

distributed electronically by NBRF (and later by the Protein Information Resource, or PIR) on

magnetic tape, and the distribution included some basic programs that could be used to search

and evaluate distant evolutionary relationships.

The next phase in the history of sequence databases was precipitated by the veritable explo-

sion in the amount of nucleotide sequence data available to researchers by the end of the

1970s. To address the need for more robust public sequence databases, the Los Alamos National

Laboratory (LANL) created the Los Alamos DNA Sequence Database in 1979, which became

known as GenBank in 1982 (Benson et al. 2018). Meanwhile, the European Molecular Biology

Laboratory (EMBL) created the EMBL Nucleotide Sequence Data Library in 1980. Throughout

the 1980s, EMBL (then based in Heidelberg, Germany), LANL, and (later) the National Center

for Biotechnology Information (NCBI, part of the National Library of Medicine at the National

Institutes of Health) jointly contributed DNA sequence data to these databases. This was done

Biological Sequence Databases

by having teams of curators manually transcribing and interpreting what was published in

print journals to an electronic format more appropriate for computational analyses. The DNA

Databank of Japan (DDBJ; Kodama et al. 2018) joined this DNA data-collecting collabora-

tion a few years later. By the late 1980s, the quantity of DNA sequence data being produced

was so overwhelming that print journals began asking scientists to electronically submit their

DNA sequences directly to these databases, rather than publishing them in printed journals

or papers. In 1988, after a meeting of these three groups (now referred to as the International

Nucleotide Sequence Database Collaboration, or INSDC; Karsch-Mizrachi et al. 2018), there

was an agreement to use a common data exchange format and to have each database update

only the records that were directly submitted to it. Thanks to this agreement, all three centers

(EMBL, DDBJ, and NCBI) now collect direct DNA sequence submissions and distribute them

so that each center has copies of all of the sequences, with each center acting as a primary distri-

bution center for these sequences. DDBJ/EMBL/GenBank records are updated automatically

every 24 hours at all three sites, meaning that all sequences can be found within DDBJ, the

European Nucleotide Archive (ENA; Silvester et al. 2018), and GenBank in short order. That

said, each database within the INSDC has the freedom to display and annotate the sequence

data as it sees fit.

In parallel with the early work being done on DNA sequence databases, the foundations

for the Swiss-Prot protein sequence database were also being laid in the early 1980s by Amos

Bairoch, recounting its history from an engaging perspective in a first-person review (Bairoch

2000). Bairoch converted PIR’s Atlas to a format similar to that used by EMBL for its nucleotide

database. In this initial release, called PIR+, additional information about each of the pro-

teins was added, increasing its value as a curated, well-annotated source of information on

proteins. In the summer of 1986, Bairoch began distributing PIR+ on the US BIONET (a pre-

cursor to the Internet), renaming it Swiss-Prot. At that time, it contained the grand sum of

3900 protein sequences. This was seen as an overwhelming amount of data, in stark contrast

to today’s standards. As Swiss-Prot and EMBL followed similar formats, a natural collaboration

developed between these two groups, and these collaborative efforts strengthened when both

EMBL’s and Swiss-Prot’s operations were moved to EMBL’s European Bioinformatics Insti-

tute (EBI; Cook et al. 2018) in Hinxton, UK. One of the first collaborative projects undertaken

by the Swiss-Prot and EMBL teams was to create a new and much larger protein sequence

database supplement to Swiss-Prot. As maintaining the high quality of Swiss-Prot entries was a

time-consuming process involving extensive sequence analysis and detailed curation by expert

annotators (Apweiler 2001), and to allow the quick release of protein data not yet annotated

to Swiss-Prot’s stringent standards, a new database called TrEMBL (for “translation of EMBL

nucleotide sequences”) was created. This supplement to Swiss-Prot initially consisted of com-

putationally annotated sequence entries derived from the translation of all coding sequences

(CDSs) found in INSDC databases. In 2002, a new effort involving the Swiss Institute of Bioin-

formatics, EMBL-EBI, and PIR was launched, called the UniProt consortium (UniProt Con-

sortium 2017). This effort gave rise to the UniProt Knowledgebase (UniProtKB), consisting

of Swiss-Prot, TrEMBL, and PIR. A similar effort also gave rise to the NCBI Protein Database,

bringing together data from numerous sources and described more fully in the text that follows.

The completion of human genome sequencing and the sequencing of numerous model

genomes, as well as the existence of a gargantuan number of sequences in general, provides

a golden opportunity for biological scientists, owing to the inherent value of these data. At

the same time, the sheer magnitude of data also presents a conundrum to the inexperienced

user, resulting not just from the size of the “sequence information space” but from the

fact that the information space continues to get larger by leaps and bounds. Indeed, the

sequencing landscape has changed significantly in recent years with the development of new

high-throughput technologies that generate more and more sequence data in a way that is

best described as “better, cheaper, faster,” with these advances feeding into the “insatiable

appetite” that scientists have for more and more sequence data (Green et al. 2017). Given the

inherent value of the data contained within these sequence databases, this chapter will focus

Nucleotide Sequence Flatfiles: A Dissection

on providing the reader with a solid understanding of these major public sequence databases,

as a first step toward being able to perform robust and accurate bioinformatic analyses.

===== PDF page 23 / printed page 3 =====

Nucleotide Sequence Databases

As described above, the major sources of nucleotide sequence data are the databases involved

in INSDC – DDBJ, ENA, and GenBank – with new or updated data being shared between

these three entities once every 24 hours. This transfer is facilitated by the use of common data

formats for the kinds of information described in detail below.

The elementary format underlying the information held in sequence databases is a text file

called the flatfile. The correspondence between individual flatfile formats greatly facilitates the

daily exchange of data between each of these databases. In most cases, fields can be mapped

on a one-to-one basis from one flatfile format to the other. Over time, various file formats have

been adopted and have found continued widespread use; others have fallen to the wayside for

a variety of reasons. The success of a given format depends on its usefulness in a variety of

contexts, as well as its power in effectively containing and representing the types of biological

data that need to be archived and communicated to scientists.

In its simplest form, a sequence record can be represented as a string of nucleotides with

some basic tag or identifier. The most widely used of these simple formats is FASTA, origi-

nally introduced as part of the FASTA software suite developed by Lipman and Pearson (1985)

that is described in detail in Chapter 3. This inherently simple format provides an easy way of

handling primary data for both humans and computers, taking the following form.

>U54469.1

CGGTTGCTTGGGTTTTATAACATCAGTCAGTGACAGGCATTTCCAGAGTTGCCCTGTTCAACAATCGATA

GCTGCCTTTGGCCACCAAAATCCCAAACTTAATTAAAGAATTAAATAATTCGAATAATAATTAAGCCCAG

TAACCTACGCAGCTTGAGTGCGTAACCGATATCTAGTATACATTTCGATACATCGAAATCATGGTAGTGT

TGGAGACGGAGAAGGTAAGACGATGATAGACGGCGAGCCGCATGGGTTCGATTTGCGCTGAGCCGTGGCA

GGGAACAACAAAAACAGGGTTGTTGCACAAGAGGGGAGGCGATAGTCGAGCGGAAAAGAGTGCAGTTGGC

For brevity, only the first few lines of the sequence are shown. In the simplest incarna-

tion of the FASTA format, the “greater than” character (>) designates the beginning of a new

sequence record; this line is referred to as the definition line (commonly called the “def line”).

A unique identifier – in this case, the accession.version number (U54469.1) – is followed by the

nucleotide sequence, in either uppercase or lowercase letters, usually with 60 characters per

line. The accession number is the number that is always associated with this sequence (and

should be cited in publications), while the version number suffix allows users to easily deter-

mine whether they are looking at the most up-to-date record for a particular sequence. The

version number suffix is incremented by one each time the sequence is updated.

Additional information can be included on the definition line to make this simple format a

bit more informative, as follows.

>ENA|U54469|U54469.1 Drosophila melanogaster eukaryotic initiation factor 4E (eIF4E)

gene, complete cds, alternatively spliced.

This modified FASTA definition line now has information on the source database (ENA),

its accession.version number (U54469.1), and a short description of what biological entity is

represented by the sequence.

合并说明：本文件由以下 fragment 按原文顺序合并而来：

• 03_Nucleotide_Sequence_Flatfiles_A_Dissection（父小节开头）
• 04_The_Header（父小节内部子标题）
• 05_The_Feature_Table（父小节内部子标题，包含 feature table 与 sequence itself）

===== PDF page 23 / printed page 3 =====

Nucleotide Sequence Flatfiles: A Dissection

As flatfiles represent the elementary unit of information within sequence databases and facil-

itate the interchange of information between these databases, it is important to understand

===== PDF page 24 / printed page 4 =====

what each individual field within the flatfile represents and what kinds of information can be

found in varying parts of the record. While there are minor differences in flatfile formats, they

can all be separated into three major parts: the header, containing information and descrip-

tors pertaining to the entire record; the feature table, which provides relevant annotations to

the sequence; and the sequence itself.

===== PDF page 24 / printed page 4 =====

The Header

The header is the most database-specific part of the record. Here, we will use the ENA version

of the record for discussion (shown in its entirety in Appendix 1.1), with the corresponding

DDBJ and GenBank versions of the header appearing in Appendix 1.2. The first line of the

record provides basic identifying information about the sequence contained in the record,

appropriately named the ID line; this corresponds to the LOCUS line in DDBJ/GenBank.

ID

U54469; SV 1; linear; genomic DNA; STD; INV; 2881 BP.

The accession number is shown on the ID line, followed by its sequence version (here, the

first version, or SV 1). As this is SV 1, this is equivalent to writing U54469.1, as described above.

This is then followed by the topology of the DNA molecule (linear) and the molecule type

(genomic DNA). The next element represents the ENA data class for this sequence (STD,

denoting a “standard” annotated and assembled sequence). Data classes are used to group

sequence records within functional divisions, enabling users to query specific subsets of the

database. A description of these functional divisions can be found in Box 1.1. Finally, the ID

line presents the taxonomic division for the sequence of interest (INV, for invertebrate; see

Internet Resources) and its length (2881 base pairs). The accession number will also be shown

separately on the AC line that immediately follows the ID lines.

Box 1.1 Functional Divisions in Nucleotide Databases

The organization of nucleotide sequence records into discrete functional types provides

a way for users to query specific subsets of the records within these databases. In addi-

tion, knowledge that a particular sequence is from a given technique-oriented database

allows users to interpret the data from the proper biological point of view. Several of these

divisions are described below, and examples of each of these functional divisions (called

“data classes” by ENA) can be found by following the example links listed on the ENA Data

Formats page listed in the Internet Resources section of this chapter.

CON

Constructed (or “contigged”) records of chromosomes, genomes, and other long DNA

sequences resulting from whole -genome sequencing efforts. The records in this

division do not contain sequence data; rather, they contain instructions for the

assembly of sequence data found within multiple database records.

EST

Expressed Sequence Tags. These records contain short (300–500 bp) single reads

from mRNA (cDNA) that are usually produced in large numbers. ESTs represent a

snapshot of what is expressed in a given tissue or at a given developmental stage.

They represent tags – some coding, some not – of expression for a given cDNA library.

GSS

Genome Survey Sequences. Similar to the EST division, except that the sequences are

genomic in origin. The GSS division contains (but is not limited to) single-pass read

genome survey sequences, bacterial artificial chromosome (BAC) or yeast artificial

chromosome (YAC) ends, exon-trapped genomic sequences, and Alu polymerase chain

reaction (PCR) sequences.

HTG

High-Throughput Genome sequences. Unfinished DNA sequences generated by

high-throughput sequencing centers, made available in an expedited fashion to the

scientific community for homology and similarity searches. Entries in this division

contain keywords indicating its phase within the sequencing process. Once finished,

HTG sequences are moved into the appropriate database taxonomic division.

===== PDF page 25 / printed page 5 =====

Nucleotide Sequence Flatfiles: A Dissection

STD

A record containing a standard, annotated, and assembled sequence.

STS

Sequence-Tagged Sites. Short (200–500 bp) operationally unique sequences that

identify a combination of primer pairs used in a PCR assay, generating a reagent that

maps to a single position within the genome. The STS division is intended to facilitate

cross-comparison of STSs with sequences in other divisions for the purpose of

correlating map positions of anonymous sequences with known genes.

WGS

Whole-Genome Shotgun sequences. Sequence data from projects using shotgun

approaches that generate large numbers of short sequence reads that can then be

assembled by computer algorithms into sequence contigs, higher -order scaffolds, and

sometimes into near-chromosome- or chromosome-length sequences.

Following the ID line are one or more date lines (denoted by DT), indicating when the entry

was first created or last updated. For our sequence of interest, the entry was originally created

on May 19, 1996 and was last updated in ENA on June 23, 2017:

DT

19-MAY-1996 (Rel. 47, Created)

DT

23-JUN-2017 (Rel. 133, Last updated, Version 5)

The release number in each line indicates the first quarterly release made after the entry

was created or last updated. The version number for the entry appears on the second line and

allows the user to determine easily whether they are looking at the most up-to-date record

for a particular sequence. Please note that this is different from the accession.version format

described above – while some element of the record may have changed, the sequence may have

remained the same, so these two different types of version numbers may not always correspond

to one another.

The next part of the header contains the definition lines, providing a succinct description

of the kinds of biological information contained within the record. The definition line (DE in

ENA, DEFINITION in DDBJ/GenBank) takes the following form.

DE

Drosophila melanogaster eukaryotic initiation factor 4E (eIF4E) gene,

DE

complete cds, alternatively spliced.

Much care is taken in the generation of these definition lines and, although many of them

can be generated automatically from other parts of the record, they are reviewed to ensure

that consistency and richness of information are maintained. Obviously, it is quite impossible

to capture all of the biology underlying a sequence in a single line of text, but that wealth of

information will follow soon enough in downstream parts of the same record.

Continuing down the flatfile record, one finds the full taxonomic information on the

sequence of interest. The OS line (or SOURCE line in DDBJ/GenBank) provides the preferred

scientific name from which the sequence was derived, followed by the common name of the

organism in parentheses. The OC lines (or ORGANISM lines in DDBJ/GenBank) contain

the complete taxonomic classification of the source organism. The classification is listed

top-down, as nodes in a taxonomic tree, with the most general grouping (Eukaryota) given

first.

OS

Drosophila melanogaster (fruit fly)

OC

Eukaryota; Metazoa; Ecdysozoa; Arthropoda; Hexapoda; Insecta; Pterygota;

OC

Neoptera; Holometabola; Diptera; Brachycera; Muscomorpha; Ephydroidea;

OC

Drosophilidae; Drosophila; Sophophora.

Each record must have at least one reference or citation, noted within what are called refer-

ence blocks. These reference blocks offer scientific credit and set a context explaining why this

particular sequence was determined. The reference blocks take the following form.

===== PDF page 26 / printed page 6 =====

Biological Sequence Databases

RN

[1]

RP

1-2881

RX

DOI; .1074/jbc.271.27.16393.

RX

PUBMED; 8663200.

RA

Lavoie C.A., Lachance P.E., Sonenberg N., Lasko P.;

RT

"Alternatively spliced transcripts from the Drosophila eIF4E gene produce

RT

two different Cap-binding proteins";

RL

J Biol Chem 271(27):16393-16398(1996).

XX

RN

[2]

RP

1-2881

RA

Lasko P.F.;

RT

;

RL

Submitted (09-APR-1996) to the INSDC.

RL

Paul F. Lasko, Biology, McGill University, 1205 Avenue Docteur Penfield,

RL

Montreal, QC H3A 1B1, Canada

In this case, two references are shown, one referring to a published paper and the other

referring to the submission of the sequence record itself. In the example above, the second

block provides information on the senior author of the paper listed in the first block, as well

as the author’s postal address. While the date shown in the second block indicates when the

sequence (and accompanying information) was submitted to the database, it does not indicate

when the record was first made public, so no inferences or claims based on first public release

can be made based on this date. Additional submitter blocks may be added to the record each

time the sequence is updated.

Some headers may contain COMMENT (DDBJ/GenBank) or CC (ENA) lines. These lines

can include a great variety of notes and comments (descriptors) that refer to the entire

record. Often, genome centers will use these lines to provide contact information and to

confer acknowledgments. Comments also may include the history of the sequence. If the

sequence of a particular record is updated, the comment will contain a pointer to the previous

versions of the record. Alternatively, if an earlier version of the record is retrieved, the

comment will point forward to the newer version, as well as backwards, if there was a still

earlier version. Finally, there are database cross-reference lines (marked DR) that provide

links to allied databases containing information related to the sequence of interest. Here, a

cross-reference to FlyBase can be seen in the complete header for this record in Appendix 1.1.

Note that the corresponding DDBJ/GenBank header in Appendix 1.2 does not contain these

cross-references.

===== PDF page 26 / printed page 6 =====

The Feature Table

Early on in the collaboration between INSDC partner organizations, an effort was made to

come up with a common way to represent the biological information found within a given

database record. This common representation is called the feature table, consisting of feature

keys (a single word or abbreviation indicating the described biological property), location infor-

mation denoting where the feature is located within the sequence, and additional qualifiers

providing additional descriptive information about the feature. The online INSDC feature table

documentation is extensive and describes in great detail what features are allowed and what

qualifiers can be used with each individual feature. Wording within the feature table uses com-

mon biological research terminology wherever possible and is consistent between DDBJ, ENA,

and GenBank entries.

Here, we will dissect the feature table for the eukaryotic transcription factor 4E gene from

Drosophila melanogaster, shown in its entirety in both Appendices 1.3 (in ENA format) and

1.4 (in DDBJ/GenBank format). This particular sequence is alternatively spliced, producing

two distinct gene products, 4E-I and 4E-II. The first block of information in the feature table is

always the source feature, indicating the biological source of the sequence and additional infor-

mation relating to the entire sequence. This feature must be present in all INSDC entries, as all

DNA or RNA sequences derive from some specific biological source, including synthetic DNA.

===== PDF page 27 / printed page 7 =====

Nucleotide Sequence Flatfiles: A Dissection

FT

source

1..2881

FT

/organism="Drosophila melanogaster"

FT

/chromosome="3"

FT

/map="67A8-B2"

FT

/mol_type="genomic DNA"

FT

/db_xref="taxon:7227"

FT

gene

80..2881

FT

/gene="eIF4E"

In the first line of the source key, notice that the numbering scheme shows the range of

positions covered by this feature key as two numbers separated by two dots (1..2881). As

the source key pertains to the entire sequence, we can infer that the sequence described in

this entry is 2881 nucleotides in length. The various ways in which the location of any given

feature can be indicated are shown in Table 1.1, accounting for a wide range of biological

scenarios. The qualifiers then follow, each preceded by a slash. The full scientific name of

the organism is provided, as are specific mapping coordinates, indicating that this sequence

is at map location 67A8-B2 on chromosome 3. Also indicated is the type of molecule that

was sequenced (genomic DNA). Finally, the last line indicates a database cross-reference

(abbreviated as db_xref) to the NCBI taxonomy database, where taxon 7227 corresponds to

D. melanogaster. In general, these cross-references are controlled qualifiers that allow entries

to be connected to an external database, using an identifier that is unique to that external

database. Following the source block above is the gene feature, indicating that the gene

itself is a subset of the entire sequence in this entry, starting at position 80 and ending at

position 2881.

FT

mRNA

join(80..224,892..1458,1550..1920,1986..2085,2317..2404,

FT

2466..2881)

FT

/gene="eIF4E"

FT

/product="eukaryotic initiation factor 4E-I"

FT

mRNA

join(80..224,1550..1920,1986..2085,2317..2404,2466..2881)

FT

/gene="eIF4E"

FT

/product="eukaryotic initiation factor 4E-II"

Table 1.1 Indicating locations within the feature table.

Single position within the sequence

345..500

A continuous range of positions bounded by and including the

indicated positions

<345..500

A continuous range of positions, where the exact lower boundary

is not known; the feature begins somewhere prior to position 345

but ends at position 500

345..>500

A continuous range of positions, where the exact upper boundary

is not known; the feature begins at position 345 but ends

somewhere after position 500

<1..888

The feature starts before the first sequenced base and continues to

position 888

(102.110)

Indicates that the exact location is unknown, but that it is one of

the positions between 102 and 110, inclusive

123 ̂ 124

Points to a site between positions 123 and 124

123 ̂ 177

Points to a site between two adjacent nucleotides or amino acids

anywhere between positions 123 and 177

join(12..78,134..202)

Regions 12–78 and 134–202 are joined to form one contiguous

sequence

complement(4918..5126)

The sequence complementary to that found from 4918 to 5126 in

the sequence record

J00194:100..202

Positions 100–202, inclusive, in the entry in this database having

accession number J00194

===== PDF page 28 / printed page 8 =====

Biological Sequence Databases

The next feature in this example indicates which regions form the two mRNA transcripts for

this gene, the first for eukaryotic initiation factor 4E-I and the second for eukaryotic initiation

factor 4E-II. In the first case (shown above), the join line indicates that six distinct DNA

segments are transcribed to form the mature RNA transcript while, in the second case, the

second region is missing, with only five distinct DNA segments transcribed into the mature

RNA transcript – hence the two splice variants that are ultimately encoded by this molecule.

FT

CDS

join(201..224,1550..1920,1986..2085,2317..2404,2466..2629)

FT

/codon_start=1

FT

/gene="eIF4E"

FT

/product="eukaryotic initiation factor 4E-II"

FT

/note="Method: conceptual translation with partial peptide

FT

sequencing"

FT

/db_xref="GOA:P48598"

FT

/db_xref="InterPro:IPR001040"

FT

/db_xref="InterPro:IPR019770"

FT

/db_xref="InterPro:IPR023398"

FT

/db_xref="PDB:4AXG"

FT

/db_xref="PDB:4UE8"

FT

/db_xref="PDB:4UE9"

FT

/db_xref="PDB:4UEA"

FT

/db_xref="PDB:4UEB"

FT

/db_xref="PDB:4UEC"

FT

/db_xref="PDB:5ABU"

FT

/db_xref="PDB:5ABV"

FT

/db_xref="PDB:5T47"

FT

/db_xref="PDB:5T48"

FT

/db_xref="UniProtKB/Swiss-Prot:P48598"

FT

/protein_id="AAC03524.1"

FT

/translation="MVVLETEKTSAPSTEQGRPEPPTSAAAPAEAKDVKPKEDPQETGE

FT

PAGNTATTTAPAGDDAVRTEHLYKHPLMNVWTLWYLENDRSKSWEDMQNEITSFDTVED

FT

FWSLYNHIKPPSEIKLGSDYSLFKKNIRPMWEDAANKQGGRWVITLNKSSKTDLDNLWL

FT

DVLLCLIGEAFDHSDQICGAVINIRGKSNKISIWTADGNNEEAALEIGHKLRDALRLGR

FT

NNSLQYQLHKDTMVKQGSNVKSIYTL"

Following the mRNA feature is the CDS feature shown above, describing the region that

ultimately encodes the protein product. Focusing just on eukaryotic initiation factor 4E-II, the

CDS feature also shows a join line with coordinates that are slightly different from those

shown in the mRNA feature, specifically at the beginning and end positions. The difference

lies in the fact that the 5′ and 3′ untranslated regions (UTRs) are included in the mRNA fea-

ture but not in the CDS feature. The CDS feature corresponds to the sequence of amino acids

found in the translated protein product whose sequence is shown in the /translation qual-

ifier above. The /codon_start qualifier indicates that the amino acid translation of the first

codon begins at the first position of this joined region, with no offset.

The /protein_id qualifier shows the accession number for the corresponding entry in

the protein databases (AAC03524.1) and is hyperlinked, enabling the user to go directly to

that entry. These unique identifiers use a “3 + 5” format – three letters, followed by five num-

bers. Versions are indicated by the decimal that follows; when the protein sequence in the

record changes, the version is incremented by one. The assignment of a gene product or pro-

tein name (via the /protein qualifier) often is subjective, sometimes being assigned via weak

similarities to other (and sometimes poorly annotated) sequences. Given the potential for the

transitive propagation of poor annotations (that is, bad data tend to beget more bad data),

users are advised to consult curated nucleotide and protein sequence databases for the most

up-to-date, accurate information regarding the putative function of a given sequence. Finally,

notice the extensive cross-referencing via the /db_xref qualifier to entries in InterPro, the

===== PDF page 29 / printed page 9 =====

Nucleotide Sequence Flatfiles: A Dissection

Protein Data Bank (PDB), and UniProtKB/Swiss-Prot, as well as to a Gene Ontology annotation

(GOA; Gene Ontology Consortium 2017).

Implicit in the source feature and the organism that is assigned to it is the genetic code used

to translate the nucleic acid sequence into a protein sequence when a CDS feature is present

in the record. Also, the DNA-centric nature of these feature tables means that all features are

mapped through a DNA coordinate system, not that of amino acid reference points, as shown

in the examples in Appendices 1.3 and 1.4.

SQ

Sequence 2881 BP; 849 A; 699 C; 585 G; 748 T; 0 other;

cggttgcttg ggttttataa catcagtcag tgacaggcat ttccagagtt gccctgttca

acaatcgata gctgcctttg gccaccaaaa tcccaaactt aattaaagaa ttaaataatt

cgaataataa ttaagcccag taacctacgc agcttgagtg cgtaaccgat atctagtata

.

. <truncated for brevity>

.

aaacggaacc ccctttgtta tcaaaaatcg gcataatata aaatctatcc gctttttgta

2820

gtcactgtca ataatggatt agacggaaaa gtatattaat aaaaacctac attaaaaccg

2880

g

2881

//

Finally, at the end of every nucleotide sequence record, one finds the actual nucleotide

sequence, with 60 bases per row. Note that, in the SQ line signaling the beginning of this section

of the record, not only is the overall length of the sequence provided, but a count of how many

of each individual type of nucleotide base is also provided, making it quite easy to compute the

GC content of this sequence.

===== PDF page 29 / printed page 9 =====

Graphical Interfaces

Graphical interfaces have been developed to facilitate the interpretation of the data found

within text-based flatfiles, with an example of the graphical view of the ENA record for our

sequence of interest (U54469.1) shown in Figure 1.1. These graphical views are particularly

useful when there is a long list of documented biological features within the feature table,

enabling the user to visualize potential interactions or relationships between biological

features. An additional example of the use of graphical views to assist in the interpretation

of the information found within a database record is provided in the discussion of the NCBI

Entrez discovery pathway in Chapter 2, as well as later in this chapter.

===== PDF page 30 / printed page 10, Figure 1.1 caption =====

Figure 1.1 The landing page for ENA record U54469.1, providing a graphical view of biological features found within the sequence of the

Drosophila melanogaster eukaryotic initiation factor 4E (eIF4E) gene. The tracks within the graphical view show the position of the gene,

mRNAs, and coding regions (marked CDS) within the 2881 bp sequence reported in this record.

===== PDF page 30 / printed page 10 =====

Box 1.2 RefSeq

The first several chapters of this book describe a variety of ways in which sequence data

and sequence annotations find their way into public databases. While the combination of

data derived from systematic sequencing projects and individual investigators’ laborato-

ries yields a rich and highly valuable set of sequence data, some problems are apparent.

The most important issue is that a single biological entity may be represented by many

different entries in various databases. It also may not be clear whether a given sequence

has been experimentally determined or is simply the result of a computational prediction.

To address these issues, NCBI developed the RefSeq project, the major goal of which

is to provide a reference sequence for each molecule in the central dogma (DNA, mRNA,

and protein). As each biological entity is represented only once, RefSeq is, by definition,

non-redundant. Nucleotide and protein sequences in RefSeq are explicitly linked to one

===== PDF page 31 / printed page 11 =====

===== PDF page 31 / printed page 11 =====

Protein Sequence Databases

With the availability of myriad complete genome sequences from both prokaryotes and eukary-

otes, significant effort is being dedicated to the identification and functional analysis of the

proteins encoded by these genomes. The large-scale analysis of these proteins continues to

generate huge amounts of data, including through the use of proteomic methods (Chapter 11)

and through protein structure analysis (Chapter 12), to name a few. These and other meth-

ods make it possible to identify large numbers of proteins quickly, to map their interactions

(Chapter 13), to determine their location within the cell, and to analyze their biological activi-

ties. This ever-increasing “information space” reinforces the central role that protein sequence

databases play as a resource for storing data generated by these efforts, making them freely

available to the life sciences community.

As most sequence data in protein databases are derived from the translation of nucleotide

sequences, they can be, in large part, thought of as “secondary databases.” Universal protein

sequence databases cover proteins from all species, whereas specialized protein sequence

databases concentrate on particular protein families, groups of proteins, or those from a

specific organism. Representative model organism databases include the Mouse Genome

Database (MGD; Smith et al. 2018) and WormBase (Lee et al. 2018), among others (Baxe-

vanis and Bateman 2015; Rigden and Fernández 2018). Organismal sequence databases are

discussed in greater detail in Chapter 2.

Universal protein databases can be divided further into two broad categories: sequence

repositories, where the data are stored with little or no manual intervention, and curated

databases, in which experts enhance the original data through expert biocuration. The

importance of ensuring interoperability, creating and implementing standards, and adopting

best practices aimed at accurately representing the biological knowledge found within the

sequence databases is absolutely paramount. Indeed, these curation goals are so important

that there is an organization called the International Society for Biocuration, the primary

mission of which is to advance these central tenets.

===== PDF page 32 / printed page 12 =====

The NCBI Protein Database

NCBI maintains the Protein database, which derives its content from a number of sources.

These include the translations of the annotated coding regions from INSDC databases

described above, from RefSeq (Box 1.2), and from NCBI’s Third Party Annotation (TPA)

database. The TPA dataset is quite interesting in its own right, as it captures both experimen-

tal and inferential data provided by the scientific community to supplement the information

found in an INSDC nucleotide entry. As the name suggests, the information in the TPA is pro-

vided by third parties and not by the original submitter of the corresponding INSDC entry. The

NCBI Protein database also includes additional non-NCBI sources of protein sequence data,

including Swiss-Prot, PIR, PDB, and the Protein Research Foundation. Step-by-step methods

for performing searches against the NCBI Protein database are described in detail in Chapter 3.

边界说明：本小节截止于下一真实小节标题 UniProt 前；PDF 同页紧随的 UniProt 标题和 Figure 1.2 图注归入后续 UniProt 小节。

===== PDF page 32 / printed page 12 =====

UniProt

Although data repositories are an essential vehicle through which scientists can access

sequence data as quickly as possible, it is clear that the addition of biological information from

Figure 1.2 Results of a search for the human heterogeneous nuclear ribosomal protein A1 record within UniProtKB, using the accession

number P09651 as the search term. See text for details.

===== PDF page 33 / printed page 13 =====

Protein Sequence Databases

multiple, highly regarded sources greatly increases the power of the underlying sequence

data. The UniProt Consortium was formed to accomplish just that, bringing together the

Swiss-Prot, TrEMBL, and the Protein Information Resource Protein Sequence Database

under a single umbrella, called UniProt (UniProt Consortium 2017). UniProt comprises

three main databases: the UniProt Archive, a non-redundant set of all publicly available

protein sequences compiled from a variety of source databases; UniProtKB, combining entries

from UniProtKB/Swiss-Prot and UniProtKB/TrEMBL; and the UniProt Reference Clusters

(UniRef), containing non-redundant views of the data contained in UniParc and UniProtKB

that are clustered at three different levels of sequence identity (Suzek et al. 2015).

The wealth of information found within a UniProtKB entry can be best illustrated by an

example. Here, we will consider the entry for the human heterogeneous nuclear ribonuclear

protein A1, with accession number P09651. A search of UniProtKB using this accession num-

ber as the search term produces the view seen in Figure 1.2. The lower part of the left-hand

column shows the various types of information available for this protein, and the user can

select or de-select sections based on their interests. The main part of the window provides basic

Figure 1.3 The Subcellular location and Pathology & Biotech sections of the record for the human heterogeneous nuclear ribosomal

protein A1 record within UniProtKB. These sections can be accessed by clicking on the blue tiles in the left-hand column of the window.

See text for details.

===== PDF page 34 / printed page 14 =====

Biological Sequence Databases

identifying information about this sequence, as well as an indication of whether the entry has

been manually reviewed and annotated by UniProtKB curators. Here, we see that the entry

has indeed been reviewed and that there is experimental evidence that supports the existence

of the protein. The next section in the file is devoted to conveying functional information, also

providing Gene Ontology (GO) terms that are associated with the entry, as well as links to

enzyme and pathway databases such as Reactome (see Chapter 13). Clicking on any of the

blue tiles in the left-hand column will jump the user down to the selected section of the entry.

For instance, if one clicks on Subcellular location, the view seen in Figure 1.3 is produced,

providing a color-coded schematic of the cell indicating the type of annotation (manual or

automatic) and links to publications supporting the annotation. The lower part of Figure 1.3

also shows information regarding the protein’s involvement in disease, documenting variants

that have been implicated in early onset Paget disease and amyotrophic lateral sclerosis (Kim

et al. 2013; Liu et al. 2016).

In the upper left corner of the UniProtKB window are display options that are quite useful

in visualizing the significant amount of data found in this entry’s feature table. By clicking

on Feature viewer, one is presented with the view shown in Figure 1.4, neatly summarizing

Figure 1.4 The Feature viewer rendering of the record for the human heterogeneous nuclear ribosomal protein A1 within UniProtKB.

Clicking the Display link, found in the upper left portion of the window, provides access to the Feature viewer. Any of the sections can be

expanded by clicking on the labels in the blue boxes to the left of the graphic. See text for details.

===== PDF page 35 / printed page 15 =====

Protein Sequence Databases

the annotations for this sequence in a coordinate-based fashion. Any of the sections can

be expanded by clicking on the labels in the blue boxes to the left of the graphic. Here, the

post-translational modification (PTM) section has been expanded, showing the position of

modified residues in this protein; clicking on any of the markers in the track will produce a

pop-up with additional information on the PTM, along with relevant links to the literature.

In Figure 1.5, the Structural features and Variants sections have also been expanded, showing

the positions of all alpha helices, beta strands, and beta turns within the protein, as well as

the location of putatively clinically relevant point mutations. Here, a variant at position 351 is

highlighted, with the proline-to-leucine variant identified as part of the ClinVar project (Lan-

drum et al. 2016) having a possible association with relapsing–remitting multiple sclerosis. By

examining different sections of this very useful graphical display, the user can start to see how

various features overlap with one another, perhaps indicating whether a known or predicted

disease-causing variant falls within a structured region of the protein. These annotations

and observations can provide important insights with respect to experimental design and the

interpretation of experimental data.

Figure 1.5 Expanding the PTM, Structural features, and Variants sections within the Feature viewer display shows the position of all

post-translational modifications (PTMs), alpha helices, beta strands, and beta turns within the human heterogeneous nuclear ribosomal

protein A1, as well as the location of putatively clinically relevant point mutations. Clicking on any of the variants produces a pop-up

window with additional information; here, the pop-up window provides disease association data for the proline-to-leucine variant at

position 351 of the sequence. See text for details.

===== PDF page 36 / printed page 16 =====

Summary

The rapid pace of discovery in the genomic and proteomic arenas requires that databases are

built in a way that facilitates not just the storage of these data, but the efficient handling

and retrieval of information from these databases. Many lessons have been learned over the

decades regarding how to approach critical questions regarding design and content, often the

hard way. Thus, the continued development of currently existing databases, as well as the

conceptualization and creation of new types of databases, will be a critical focal point for

the advancement of biological discovery. As should be obvious from this chapter, keeping

databases up to date and accurate is a task that requires the active involvement of the bio-

logical community (Box 1.3). Therefore, it is incumbent upon all users to ensure the accuracy

of these data in an active fashion, engaging the curators in a continuous dialog so that these

widely used resources continue to remain a valuable resource to biologists worldwide.

Box 1.3 Ensuring the Continued Quality of Data in Public Sequence Databases

Given the roles of DDBJ, EMBL, and GenBank in maintaining the archive of all publicly

available DNA, RNA, and protein sequences, the continued usefulness of this resource

is highly dependent on the quality of data found within it. Despite the high degree of

both manual and automated checking that takes place before a record becomes pub-

lic, errors will still find their way into the databases. These errors may be trivial and

have no biological consequence (e.g. an incorrect postal code), may be misleading (e.g.

an organism having the correct genus but wrong species name), or downright incorrect

(e.g. a full-length mRNA not having a CDS annotated on it). Sometimes, records may have

incorrect reference blocks, preventing researchers from linking to the correct publication

describing the sequence. Over time, many have taken an active role in reporting these

errors but, more often than not, these errors are left uncorrected.

While the individual INSDC members have the responsibility for hosting and dissemi-

nating the data found within their databases, keep in mind that the ownership of the data

rests with the original submitter – and these original submitters (or their designees) are

the only ones who can make updates to their database records. To keep these community

resources as accurate and up to date as possible, users are actively encouraged to report

any errors found when using the databases in the course of their work so that the database

administrators can follow up with the original submitters as appropriate.

Given below are the current e-mail addresses for submitting information regarding

errors to the three major sequence databases. As all the databases share information with

each other nightly, it is only necessary to report the error to one of the three members of

the consortium. Authors are actively encouraged to check their own records periodically

to ensure that the information they previously submitted is still accurate. Even though

this charge to the community is discussed here in the context of the three major sequence

databases, all databases provide similar mechanisms through which incorrect information

can be brought to the attention of the database administrators.

DDBJ

ddbjupdt@ddbj.nig.ac.jp

EMBL

datasubs@ebi.ac.uk

GenBank

gb-admin@ncbi.nlm.nih.gov

As alluded to above, the range of publicly available data obviously goes well beyond human

data, whether sequence based or not. As the major public sequence databases need to be able to

store data in a fairly generalized fashion, these databases often do not contain more specialized

types of information that would be of interest to specific segments of the biological community.

To address this, many smaller, specialized databases have emerged and have been developed

and curated by biologists “in the trenches” to fulfill specific needs. These databases, which

contain information ranging from strain crosses to gene expression data, provide a valuable

===== PDF page 37 / printed page 17 =====

Further Reading

adjunct to the more visible public sequence databases, and users are encouraged to make intel-

ligent use of both types of databases. An annotated list of such databases can be found in the

yearly Database issue of Nucleic Acids Research (Rigden and Fernández 2018).

The position of this chapter at the beginning of this book reflects the belief that an

understanding of biological databases is the first step toward being able to perform robust

and accurate bioinformatic analyses. The reader is very strongly encouraged to take the

time to understand the structure of the data found within these databases, as the basis for

finding sequence data of interest and performing the more advanced analyses described in

the chapters that follow.

===== PDF page 37 / printed page 17 =====

Acknowledgments

The author thanks Rolf Apweiler for the use of material from the third edition of this book.

Internet Resources

DDBJ Database Divisions

www.ddbj.nig.ac.jp/ddbj/data-categories-e.html

DNA Database of Japan (DDBJ)

www.ddbj.nig.ac.jp

EMBL Nucleotide Sequence Database

www.embl.org

ENA Data Formats

www.ebi.ac.uk/ena/submit/data-formats

European Bioinformatics Institute

www.ebi.ac.uk

GenBank

www.ncbi.nlm.nih.gov

GenBank Database Divisions

www.ncbi.nlm.nih.gov/genbank/htgs/divisions

Genome Ontology Consortium

geneontology.org

INSDC Feature Table Definition

insdc.org/documents/feature_table.html

International Society for Biocuration

biocuration.org

NCBI Data Model

www.ncbi.nlm.nih.gov/IEB/ToolBox/SDKDOCS/

DATAMODL.HTML

NCBI Protein Database

www.ncbi.nlm.nih.gov/protein

Nucleic Acids Research Database issue

academic.oup.com/nar

Protein Data Bank (PDB)

www.rcsb.org

Protein Identification Resource (PIR)

pir.georgetown.edu

Protein Research Foundation

www.proteinresearch.net

RefSeq

www.ncbi.nlm.nih.gov/refseq

Swiss-Prot (EBI)

www.ebi.ac.uk/uniprot

Swiss-Prot (ExPASy)

web.expasy.org/docs/swiss-prot_guideline.html

UniProt Consortium

www.uniprot.org

===== PDF page 37 / printed page 17 =====

Further Reading

Bairoch, A. (2000). Serendipity in bioinformatics: the tribulations of a Swiss bioinformatician

through exciting times! Bioinformatics. 16: 48–64. A personal narrative conveying the early

history of the development of sequence databases and related software tools, events that set the

groundwork for the modern bioinformatics landscape.

Green, E.D., Rubin, E.M., and Olson, M.V. (2017). The future of DNA sequencing. Nature. 550:

179–181. An insightful perspective regarding the next several decades of the application of DNA

sequencing methodologies in novel contexts and the implications of those applications to issues

of data storage and data sharing.

Rigden, D.J. and Fernández, X.M. (2018). The 2018 Nucleic Acids Research database issue and the

online molecular biology database collection. Nucleic Acids Res. 46: D1–D7. The 25th overview

===== PDF page 38 / printed page 18 =====

Biological Sequence Databases

of the annual database issue published by Nucleic Acids Research, capturing the wide variety of

publicly available bioinformatic databases available to the community. This overview is updated

yearly, and the individual papers describing these database resources are freely available

through the Nucleic Acids Research web site.

References

Apweiler,R.(2001).Functional information in Swiss-Prot: the basis for large-scale characterization of protein sequences. Briefings Bioinf. 2:9-18.

Bairoch,A.(2000).Serendipity in bioinformatics: the tribulations of a Swiss bioinformatician through exciting times! Bioinformatics. 16:48-64.

Baxevanis,A.D. and Bateman,A.(2015).The importance of biological databases in biological discovery. Curr. Protoc. Bioinf. 50:1.1.1-1.1.8.

Benson,D.A., Cavanaugh,M., Clark,K. et al.(2018).GenBank. Nucleic Acids Res. 46:D41-D47.

Cook,C.E., Bergman,M.T., Cochrane,G. et al.(2018).The European Bioinformatics Institute in 2017: data coordination and integration. Nucleic Acids Res. 46:D21-D29.

Dayhoff,M.O., Eck,R.V., Chang,M.A., and Sochard,M.R.(1965). Atlas of Protein Sequence and Structure. Silver Spring, MD: National Biomedical Research Foundation.

Gene Ontology Consortium(2017).Expansion of the Gene Ontology knowledgebase and resources. Nucleic Acids Res. 45:D331-D338.

Green,E.D., Rubin,E.M., and Olson,M.V.(2017).The future of DNA sequencing. Nature. 550:179-181.

Karsch-Mizrachi,I., Tagaki,T., and Cochrane,G., on behalf of the International Nucleotide Sequence Database Collaboration(2018).The International Nucleotide Sequence Database Collaboration. Nucleic Acids Res. 46:D48-D51.

Kim,H.J., Kim,N.C., Wang,Y.D. et al.(2013).Mutations in prion-like domains in hnRNPA2B1 and hnRNPA1 cause multisystem proteinopathy and ALS. Nature. 495:467-473.

Kodama,Y., Mashima,J., Kosuge,T. et al.(2018).DNA Data Bank of Japan: 30th anniversary. Nucleic Acids Res. 46:D30-D35.

Landrum,M.J., Lee,J.M., Benson,M. et al.(2016).ClinVar: public archive of interpretations of clinically relevant variants. Nucleic Acids Res. 44:D862-D868.

Lee,R.Y.N., Howe,K.L., Harris,T.W. et al.(2018).WormBase 2017: molting into a new stage. Nucleic Acids Res. 46:D869-D874.

Lipman,D.J. and Pearson,W.R.(1985).Rapid and sensitive protein similarity searches. Science. 227:1435-1441.

Liu,Q., Shu,S., Wang,R.R. et al.(2016).Whole-exome sequencing identifies a missense mutation in hnRNPA1 in a family with flail arm ALS. Neurology. 87:1763-1769.

Rigden,D.J. and Fernandez,X.M.(2018).The 2018 Nucleic Acids Research database issue and the online molecular biology database collection. Nucleic Acids Res. 46:D1-D7.

Silvester,N., Alako,B., Amid,C. et al.(2018).The European Nucleotide Archive in 2017. Nucleic Acids Res. 46:D36-D40.

Smith,C.L., Blake,J.A., Kadin,J.A. et al., and The Mouse Genome Database Group(2018).Mouse Genome Database (MGD)-2018: knowledgebase for the laboratory mouse. Nucleic Acids Res. 46:D836-D842.

Suzek,B.E., Wang,Y., Huang,H. et al., and The UniProt Consortium(2015).UniRef clusters: a comprehensive and scalable alternative for improving sequence similarity searches. Bioinformatics. 31:926-932.

UniProt Consortium(2017).UniProt: the universal protein knowledgebase. Nucleic Acids Res. 45:D158-D169.

章节作者声明

This chapter was written by Dr. Andreas D. Baxevanis in his private capacity. No official support or endorsement by the National Institutes of Health or the United States Department of Health and Human Services is intended or should be inferred.