Biological Databases | Access to sequence data and related information

CHAPTER 2:
‘ACCESS TO SEQUENCE DATA &
RELATED INFORMATION’
PREPARED BY:
NAHAL JEHANGIR MALIK
1

2
The body of data available in protein sequences is something fundamentally new in
biology and biochemistry, unprecedented in quantity, in concentrated information
content and in conceptual simplicity … For the past four years we have published an
annual Atlas of Protein Sequence and Structure, the latest volume of which contains
nearly 500 sequences or partial sequences established by several hundred workers in
various laboratories.
-----Margaret Dayhoff (1969)

Introduction to Biological Databases
■ All living organisms are characterized by the capacity to reproduce and evolve.
■ The genome of an organism is defined as the collection of DNA within that
organism, including the set of genes that encode RNA molecules and proteins.
■ In 1995 the complete genome of a free-living organism was sequenced for the first
time, the bacterium Haemophilus influenzae. In the years since then the genomes
of thousands of organisms have been completely sequenced, ushering in a new
era of biological data acquisition and information accessibility.
■ Publicly available databanks now contain quadrillions (>1015) of nucleotides of
DNA sequence data, soon to be quintillions (>1018 bases). These have been
collected from over 300,000 different species of organisms.
■ The goal of this chapter is to introduce the databases that store these data and
strategies to extract information from them.
3

DNA Sequencing
■ There are two main technologies for DNA sequencing.
■ Sanger Sequencing: Beginning in the 1970s dideoxynucleotide sequencing
“Sanger sequencing” was the principal method.
■ NGS: Since 2005 next-generation sequencing (NGS) technology has emerged,
allowing orders of magnitude more sequence data to be generated.
■ The availability of vastly more sequence data (at a relatively low cost per base) has
impacted most areas of bioinformatics and genomics. There are new challenges in
acquiring, analyzing, storing, and distributing such data. It is no longer unusual for
researchers to analyze datasets that are many terabytes in size.
■ In this chapter we will introduce two ways of thinking about accessing data:
1) Individual genes, proteins & related molecules
2) Large datasets related to interest problem
4

First perspective: Individual genes/proteins
■ Taking human beta globin as an example, there is a locus (on chromosome 11)
harboring the beta globin gene (HBB) and associated genomic elements such as a
promoter and introns.
■ There is tremendous variation between people (variants include single-nucleotide
variants, differences in repetitive DNA elements, and differences in chromosomal
copy number).
■ This gene can be transcribed to beta globin mRNA which is expressed in particular
tissues (and particular times of development) and may be translated into beta
globin protein.
■ This protein is a subunit of the hemoglobin protein, a tetramer that has various
functions in health and diseases. All this information about the beta globin gene,
RNA, and protein is accessible through the databases and resources introduced in
this chapter.
5

Second perspective: Large Datasets
A second perspective is on large datasets related to a problem of interest. Here are three
examples:
1. We might want to study all the variants that have been identified across all human
globin genes.
2. In patients having mutations in a gene we might want to study the collection of all of
the tens of thousands of RNA transcripts in a given cell type in order to assess the
functional consequences of that variation. After performing a microarray or RNAseq
experiment, it might be of interest to identify a set of regulated transcripts and assign
their protein products to some cellular pathways.
3. Perhaps we want to sequence the DNA corresponding to a set of 100 genes
implicated in hemoglobin function. Databases and resources such as Entrez, BioMart,
and Galaxy facilitate the manipulation of larger datasets. You can acquire, store, and
analyze datasets involving some set of molecules that have been previously
characterized (e.g., all known protein-coding genes on human chromosome 11) or are
novel (e.g., data you obtain experimentally that you can annotate and compare to
known data).
6

Centralized Databases Store DNA Sequences
How much DNA sequence is stored in public databases? Where are the data stored?
We begin with three main sites that have been responsible for storing nucleotide
sequence data from 1982 to the present These are:
(1) GenBank at the National Center for Biotechnology Information (NCBI) of the
National Institutes of Health (NIH) in Bethesda
(2) European Molecular Biology Laboratory (EMBL)-Bank Nucleotide Sequence
Database (EMBL-Bank), part of the European Nucleotide Archive (ENA) at the
European Bioinformatics Institute (EBI) in Hinxton, England
(3) DNA Database of Japan (DDBJ) at the National Institute of Genetics in Mishima
All three are coordinated by the International Nucleotide Sequence Database
Collaboration (INSDC), and they share their data daily.
GenBank, EMBLBank, and DDBJ are organized as databases within NCBI, EBI, and
DDBJ which offer many dozens of other resources for the study of sequence data.
7

9
National Center for Biotechnology Information (NCBI):
organization
FIG.2.2a

10
European Bioinformatics Institute (EBI): organization

11
DNA Database of Japan (DDBJ): organization

■ Members of the research community can submit records directly to sequence
repositories at NCBI, EBI, and DDBJ.
■ Quality control is assured through guidelines enforced at the time of submission,
and through projects such as RefSeq that reconcile differences between submitted
entries.
■ For GenBank, NCBI offers the command-line tool tbl2asn to automate the creation
of sequence records. GenBank (representative of the holdings of EMBL-Bank and
DDBJ) has received submissions since 1982, including sequences from thousands
of individual submitters.
■ Over the past 30 years the number of bases in GenBank has doubled
approximately every 18 months. GenBank, EMBL-Bank, and DDBJ accept sequence
data that consist of complete or incomplete genomes (or chromosomes) analyzed
by a whole-genome shotgun (WGS) strategy. The WGS division consists of
sequences generated by high-throughput sequencing efforts.
■ WGS sequences have been available since 2002, but they are not considered part
of the GenBank/EMBLBank/DDBJ releases.
12

FIG 2.3:
■ Inspection of Figure 2.3 reveals that the recently developed Sequence Read
Archive (SRA): contains vastly more sequence data than the sum of GenBank and
WGS; in fact, SRA currently holds 3000 times more bases of DNA.
■ Each sequence read in SRA is relatively short (typically 50–400 base pairs),
reflecting next-generation sequencing technology.
■ Most of the SRA data are publicly available (such as sequences from various
organisms across the tree of life); these are shown as open-access bases in Figure
2.3 Some of the data are derived from humans, and can potentially lead to the
identification of particular clinical subjects or research participants.
■ Access to those data is therefore restricted, requiring application to a committee
from qualified researchers who agree to adhere to ethical guidelines. Figure 2.3
shows data from SRA at NCBI, including total data and open access data.
14

Just Read Points
■ The first eukaryotic genome to be completed (Saccharomyces cerevisiae) is ∼13
million base pairs (Mb) in size. An average-sized human chromosome is ∼150 Mb, and
a single human genome consists of >3 billion base pairs (3 Gb).
■ For next-generation sequencing, short sequence reads (typically 100– 300 base pairs in
length) are obtained in vast quantities that allow each single base pair to be
represented (“covered”) by some average number of independent reads such as 30.
There is therefore 30 × depth of coverage.
■ For a recent study in my lab, we obtained paired affected/unaffected samples from
three individuals with a disease, performed whole-genome sequencing, and obtained
700 billion (7 × 1011) bases of DNA sequence.
■ For a large-scale cancer study involving 20,000 tumor/normal comparisons, a massive
1016 bases of DNA can be generated.
■ Even larger studies involving 200,000 tumor/normal comparisons are being planned.
■ Other experimental approaches such as whole-exome sequencing (involving
sequencing the collection of exons in a genome that are thought to be functionally
most important) and sequencing of the transcriptome also generate large amounts of
data. 16

We can also consider amounts of
sequence data in terms of terabytes.
A byte is a unit of computer storage
information, consisting of 8 bits and
encoding one character.
17
Size Abbreviatio
n
# bytes Example
Bytes -- 1 Single text character
Kilobytes 1 kb 103 Text file, 1000
characters
Megabytes 1 MB 106 Text file, 1m
characters
Gigabytes 1 GB 109 Size of GenBank: 600
GB
Terabytes 1 TB 1012 Size of 1000
Genomes Project:
<500 TB
Petabytes 1 PB 1015 Size of SRA at NCBI:
5 PB
Exabytes 1 EB 1018 Annual worldwide
output: >2 EB

INFORMATION:
■ A typical desktop might have 500 gigabytes (Gb) of storage. The uncompressed
flatfiles of the current release of GenBank (introduced below) are ∼600 Gb. One
thousand gigabytes is equivalent to one terabyte (1000 Gb = 1 Tb), which is the
amount of storage some researchers use to study a singe whole human genome.
■ One thousand terabytes is equivalent to one petabyte (1000 Tb = 1 Pb). Large-
scale sequencing projects, for example one that involves whole-genome sequences
of 10,000 individuals, require several Pb of storage.
■ A megabase is one million (106) bases of DNA. A gigabase is one billion (109)
bases. A terabase is one trillion (1012) bases.
18

Contents of DNA, RNA, and Protein databases
■ While the sequence information underlying DDBJ, EMBL-Bank, and GenBank are
equivalent, we begin our discussion with GenBank. GenBank is a database
consisting of most known public DNA and protein sequences, excluding next-
generation sequence data.
■ In addition to storing these sequences, GenBank contains bibliographic and
biological annotation. Its data are available free of charge from NCBI.
19

Organisms in Genbank/EMbL-bank/DDBJ
■ Over 310,000 different species are represented in GenBank, with over 1000 new
species added per month.
■ Briefly, eukaryotes have a nucleus and are often multicellular, while bacteria do not
have a nucleus.
■ Archaea are single-celled organisms, distinct from eukaryotes and bacteria, and
constitute a third major branch of life.
■ Viruses, which contain nucleic acids (DNA or RNA) but can only replicate in a host
cell, exist at the borderline of the definition of living organisms.
■ Of the microbes, there are about 25 times more bacterial than archaeal species
represented in GenBank.
20

21
TABLE 2.3: Taxa represented in GenBank (at NCBI)

TABLE 2.4:
■ The number of entries and bases of DNA/RNA for the 10 most sequenced
organisms in GenBank is provided in table 2.4 (excluding chloroplast and
mitochondrial sequences).
■ This list includes some of the most common model organisms that are studied in
biology.
■ Notably, the scientific community is studying a series of mammals (e.g., human,
mouse, cow), other vertebrates (chicken, frog), and plants (corn, rice, bread wheat,
wine grape). Different species are useful for a variety of different studies.
■ Bacteria, archaea, fungi, and viruses are absent from the list in table 2.4 because
they have relatively small genomes
22

23
TABLE2.4. Top ten organisms for which
expressed sequence tags (ESTs) have been
sequenced

FILE DIVISION
To help organize the available information, each sequence name in a GenBank record
is followed by its data file division and primary accession number. (We will define
accession numbers below.)
The following codes are used to designate the data file divisions:
1. PRI: primate sequences
2. ROD: rodent sequences
3. MAM: other mammalian sequences
4. VRT: other vertebrate sequences
5. INV: invertebrate sequences
6. PLN: plant, fungal, and algal sequences
7. BCT: bacterial sequences
8. VRL: viral sequences
24

9. PHG: bacteriophage sequences
10. SYN: synthetic sequences
11. UNA: unannotated sequences
12. EST: expressed sequence tags
13. PAT: patent sequences
14. STS: sequence-tagged sites
15. GSS: genome survey sequences
16. HTG: high-throughput genomic sequences
17. HTC: high-throughput cDNA sequences
18. ENV: environmental sampling sequences
19. CON: constricted sequences
20. TSA: transcriptome shotgun assembly sequences.
25

Types of Data in GenBank/EMBL-Bank/DDBJ
■ There are enormous numbers of molecular sequences in the DDBJ, EMBL-Bank,
and GenBank databases. We will next look at some of the basic kinds of data
present in GenBank. We then address strategies to extract the data you want from
GenBank.
■ We start with an example. We want to find out the sequence of human beta globin.
A fundamental distinction is that both DNA, RNA-based, and protein sequences are
stored in discrete databases.
■ Furthermore, within each database sequence data are represented in a variety of
forms. For example, beta globin may be described at the DNA level (e.g., as a
gene), at the RNA level (as a messenger RNA or mRNA transcript), and at the
protein level. Because RNA is relatively unstable, it is typically converted to
complementary DNA (cDNA), and a variety of databases contain cDNA sequences
corresponding to RNA transcripts.
26

■ Beginning with the DNA, a first task is to learn the official name and symbol of a
gene (and its gene products, including the protein).
■ Beta globin has the official name of “hemoglobin, beta” and the symbol HBB.
■ From one point of view there is no such thing as a “hemoglobin gene” because
globin genes encode globin proteins, and the combination of these globins with
heme forms the various types of hemoglobin. Perhaps “globin, beta” might be a
more appropriate official name.) For humans and many other species, the RNA or
cDNA is generally given the same name (e.g., HBB), while the protein name may
differ and is not italicized.
■ Often, multiple investigators study the same gene or protein and assign different
names. The human genome organization (HUGO) Gene Nomenclature Committee
(HGNC) has the critical task of assigning official names to genes and proteins.
■ For our example of beta globin, the various forms are described in the NEXT
SECTION.
27

Genomic DNA Databases
Gene:
■ A gene is localized to a chromosome. The gene is the functional unit of heredity
and is a DNA sequence that typically consists of regulatory regions, protein-coding
exons, and introns. Often, human genes are 10–100 kb in size.
■ In the case of human HBB this gene is situated on chromosome 11. The beta
globin gene may be part of a large fragment of DNA such as a cosmid, bacterial
artificial chromosome (BAC), or yeast artificial chromosome (YAC) that may
contain several genes.
■ A BAC is a large segment of DNA (typically up to 200,000 base pairs or 200 kb)
that is cloned into bacteria.
■ Similarly, YACs are used to clone large amount of DNA into yeast.
■ BACs and YACs are useful vectors with which to sequence large portions of
genomes.
29

 DNA-Level Data: Sequence-Tagged Sites
(STSs)
■ The Probe database at NCBI includes STSs, which are short (typically 500 base
pairs long) genomic landmark sequences for which both DNA sequence data and
mapping data are available.
■ STSs have been obtained from several hundred organisms, including primates and
rodents. Because they are sometimes polymorphic, containing short sequence
repeats. STSs can be useful for mapping studies.
30

 DNA-Level Data: Genome Survey
Sequences (GSSs)
■ All searches of the NCBI Nucleotide database provide results that are divided into
three sections: GSS, ESTs, and “CoreNucleotide” (i.e.,the remaining nucleotide
sequences. The GSS division of GenBank consists of sequences that are genomic
in origin (in contrast to entries in the EST division which are derived from cDNA
[mRNA]). The GSS division contains the following types of data:
 random “single-pass read” genome survey sequences
 cosmid/BAC/YAC end sequences
 exon-trapped genomic sequences
 the Alu polymerase chain reaction (PCR) sequences.
31

 DNA-Level Data: High-Throughput Genomic
Sequence (HTGS)
■ The HTGS division was created to make “unfinished” genomic sequence data
rapidly available to the scientific community. It was set up from a coordinated
effort between the three international nucleotide sequence databases: DDBJ,
EMBL, and GenBank.
■ The HTGS division contains unfinished DNA sequences generated by the high-
throughput sequencing centers
32

POINTS:
■ There are currently 38 million GSS entries from over 1000 organisms (February
2015). The top four organisms account for about onethird of all entries (these are
the mouse Mus musculus, a marine metagenome collection, the maize Zea mays,
and human).
■ Human chromosome 11, which is a mid-sized chromosome, contains about 1800
genes and is about 134 × 106 base pairs (134 Mb) in length.
■ In DNA databases, the convention is to use the four DNA nucleotides (guanine,
adenine, thymidine, cytosine; G, A, T, C) when referring to DNA derived from RNA.
The RNA base uridine (U) corresponding to T is not used.
33

RNA Data
We have described some of the basic kinds of DNA sequence data in GenBank,
EMBLBank, and DDBJ. We next consider RNA-level data.
RNA-Level Data: cDNA Databases Corresponding to Expressed Genes:
 Protein-coding genes, pseudogenes, and noncoding genes are all transcribed from
DNA to RNA.
 Genes are expressed from particular regions of the body and times of
development. If one obtains a tissue such as liver, purifies RNA, then converts the
RNA to the more stable form of complementary DNA (cDNA), some of the cDNA
clones contained in that cDNA are likely to encode beta globin.
 Beta globin RNA is therefore represented in databases as an expressed sequence
tag (EST), that is, a cDNA sequence derived from a particular cDNA library.
34

RNA-Level Data: Expressed Sequence Tags
(ESTs)
■ The database of expressed sequence tags (dbEST) is a division of GenBank that
contains sequence data and other information on “single-pass” cDNA sequences from a
number of organisms.
■ An EST is a partial DNA sequence of a cDNA clone. All cDNA clones, and therefore all
ESTs, are derived from some specific RNA source such as human brain or rat liver. The
RNA is converted into a more stable form, cDNA, which may then be packaged into a
cDNA library.
■ Typically ESTs are randomly selected cDNA clones that are sequenced on one strand
(and therefore may have a relatively high sequencing error rate). ESTs are often 300–
800 base pairs in length. The earliest efforts to sequence ESTs resulted in the
identification of many hundreds of genes that were novel at the time.
■ Currently, GenBank divides ESTs into three major categories: human, mouse, and other.
table 2.5 shows the 10 organisms from which the greatest number of ESTs has been
sequenced. Assuming that there are 20,300 human protein-coding genes and given
that there are about 8.7 million human ESTs, there is currently an average of over 400
ESTs corresponding to each human protein-coding gene.
35

RNA-Level Data: UniGene
■ The goal of the UniGene (unique gene) project is to create gene-oriented clusters by
automatically partitioning ESTs into nonredundant sets. Ultimately there should be
one UniGene cluster assigned to each gene of an organism.
■ There may be as few as one EST in a cluster, reflecting a gene that is rarely
expressed, to tens of thousands of ESTs associated with a highly expressed gene.
■ The 19 phyla containing 142 organisms currently represented in UniGene are listed
in table 2.6.
■ For human beta globin, there is only a single UniGene entry. This entry currently
has ∼2400 human ESTs that match the beta globin gene. This large number of
ESTs reflects how abundantly the beta globin gene has been expressed in cDNA
libraries that have been sequenced.
■ A UniGene cluster is a database entry for a gene containing a group of
corresponding ESTs.
36

37
UniGene database: clusters of EST
sequences
FIG2.5a

There are now thought to be approximately 20,300 human protein-coding genes. One
might expect an equal number of UniGene clusters. However, there are far more
human UniGene clusters (currently 130,000) than there are genes.
This discrepancy could occur for three reasons;
1) Much of the genome is transcribed at low levels. Currently (UniGene build 235),
64,000 human UniGene clusters consist of a single EST and ∼100,000 UniGene
clusters consist of just 1–4 ESTs. These could reflect rare transcription events of
unknown biological relevance.
2) Some DNA may be transcribed during the creation of a cDNA library without
corresponding to an authentic transcript; it is therefore a cloning artifact.
Alternative splicing (Chapter 10) may introduce apparently new clusters of genes
because the spliced exon has no homology to the rest of the sequence.
3) Clusters of ESTs could correspond to distinct regions of one gene. In that case
there would be two (or more). As a genome sequence becomes finished, it may
become apparent that the two UniGene clusters should properly cluster into one.
The number of UniGene clusters may therefore collapse over time.
SEE FIG2.5 for UNICLUSTERS
38

Protein Databases
■ In many cases you are interested in obtaining protein sequences. The Protein
database at NCBI consists of translated coding regions from GenBank as well as
sequences from external databases such as UniProt (UniProt Consortium 2012),
The Protein Information Resource (PIR), SWISS-PROT, Protein Research Foundation
(PRF), and the Protein Data Bank (PDB).
■ The EBI similarly provides information on proteins via these major databases. We
will next explore ways to obtain protein data through UniProt, an authoritative and
comprehensive protein database.
40

UniProt:
The Universal Protein Resource (UniProt) is the most comprehensive, centralized
protein sequence catalog. Formed as a collaborative effort in 2002, it consists of a
combination of three key databases;
1). Swiss-Prot is considered the best-annotated protein database, with descriptions of
protein structure and function added by expert curators.
2). The translated EMBL (TrEMBL) Nucleotide Sequence Database Library provides
automated (rather than manual) annotations of proteins not in Swiss-Prot. It was
created because of the vast number of protein sequences that have become available
through genome sequencing projects.
3). PIR maintains the Protein Sequence Database, another protein database curated
by experts.
41

■ UniProt is organized in three database layers:
1). The UniProt Knowledgebase (UniProtKB) is the central database that is divided into
the manually annotated UniProtKB/Swiss-Prot and the computationally annotated
UniProtKB/TrEMBL.
2). The UniProt Reference Clusters (UniRef) offer nonredundant reference clusters
based on UniProtKB. UniRef clusters are available with members sharing at least 50%,
90%, or 100% identity.
3). The UniProt Archive, UniParc, consists of a stable, nonredundant archive of protein
sequences from a wide variety of sources (including model organism databases,
patent offices, RefSeq, and Ensembl).
 You can access UniProt directly from its website, or from EBI or ExPASy. A search
for beta globin yields dozens of results. At present RefSeq accessions are not
displayed, so for a given query it may be unclear which sequence is the prototype.
42

Central BioInformatics Resources;
NCBI and EBI:
■ We have looked at the amount of DNA in centralized databases, and the types of
DNA, RNA, and protein entries.
■ We next visit two of the main centralized bioinformatics hubs: the National Center
for Biotechnology Information (NCBI) and the European Bioinformatics Institute
(EBI).
43

Introduction to NCBI:
The NCBI creates public databases, conducts research in computational biology,
develops software tools for analyzing genome data, and disseminates biomedical
information. Prominent resources include the following:
■ PubMed is the search service from the National Library of Medicine (NLM) that
provides access to over 24 million citations in MEDLINE (Medical Literature,
Analysis, and Retrieval System Online) and other related databases, with links to
participating online journals.
■ Entrez integrates the scientific literature, DNA, and protein sequence databases,
three-dimensional protein structure data, population study datasets, and
assemblies of complete genomes into a tightly coupled system. PubMed is the
literature component of Entrez.
■ BLAST (Basic Local Alignment Search Tool) is NCBI’s sequence similarity search
tool designed to support analysis of nucleotide and protein databases. BLAST is a
set of similarity search programs designed to explore all of the available sequence
databases, regardless of whether the query is protein or DNA.
44

■ Online Mendelian Inheritance in Man (OMIM) is a catalog of human genes
and genetic disorders. It was created by Victor McKusick and his colleagues and
developed for the World Wide Web by NCBI (Amberger et al., 2011). The database
contains detailed reference information. It also contains links to PubMed articles
and sequence information. We describe OMIM on human disease.
■ Books: NCBI offers about 200 books online. These books are searchable, and are
linked to PubMed.
■ Taxonomy: the NCBI taxonomy website includes a taxonomy browser for the
major divisions of living organisms (archaea, bacteria, eukaryota, and viruses). The
site features taxonomy information such as genetic codes and taxonomy resources
and additional information such as molecular data on extinct organisms and
recent changes to classification schemes.
45

■ Structure: the NCBI structure site maintains the Molecular Modelling Database
(MMDB), a database of macromolecular three-dimensional structures, as well as
tools for their visualization and comparative analysis. MMDB contains
experimentally determined biopolymer structures obtained from the Protein Data
Bank (PDB). Structure resources at NCBI include PDBeast (a taxonomy site within
MMDB), Cn3D (a three-dimensional structure viewer), and a vector alignment
search tool (VAST) which allows comparison of structures.
47

The European Bioinformatics Institute (EBI)
■ The EBI website is comparable to NCBI in its scope and mission, and it represents
a complemenatary, independent resource. EBI features six core molecular
databases;
1) EMBL-Bank is the repository of DNA and RNA sequences that is complementary to
GenBank and DDBJ
2) Swiss-Prot &
3) TrEMBL are two protein databases
4) MSD is a protein structure database
5) Ensembl is one of the main genome browsers; and
6) ArrayExpress is one of the two main worldwide repositories for gene expression
data, along with the Gene Expression Omnibus at NCBI.
48

■ It is often helpful to explore the wealth of resources in both these sites. For
example, each offers expert functional annotation of particular sequences and
expert curation of databases. The NCBI and EBI websites increasingly offer an
integration of their database resources so that information between the two sites
can be easily linked.
Ensembl:
Founded in 1999 to annotate the human genome, the Ensembl project now spans
over 70 vertebrate species. Related Ensembl projects include hundreds of other
species from insects to bacteria.
49

Accession Numbers:
An essential feature of DNA and protein sequence records is that they are tagged with
accession numbers.
An accession number is a string of about 4–12 numbers and/ or alphabetic
characters that are associated with a molecular sequence record (some are much
longer). An accession number may also label other entries, such as protein structures
or the results of a gene expression experiment. Accession numbers from molecules in
different databases have characteristic formats (Box 2.2). These formats vary because
each database employs its own system.
50

Accession Numbers Format
CH471100.2 GenBank genomic DNA sequence
NC_000001.10 Genomic contig
rs121434231 dbSNP (single nucleotide polymorphism)
AI687828.1 An expressed sequence tag (1 of 184)
NM_001206696 RefSeq DNA sequence (from a transcript)
NP_006138.1 RefSeq protein
CAA18545.1 GenBank protein
O14896 SwissProt protein
1KT7 Protein Data Bank structure record
51
→DNA
→RNA
→PROTEINS

■ For a typical molecule such as beta globin there are thousands of accession
numbers.
■ Many of these correspond to ESTs and other fragments of DNA that match beta
globin. How can you assess the quality of sequence or protein data? Some
sequences are full-length, while others are partial. Some reflect naturally occurring
variants such as single-nucleotide polymorphisms, or alternatively spliced
transcripts.
■ Many of the sequence entries contain errors, particularly in the ends of EST reads.
When we compare beta globin sequence derived from mRNA and from genomic
DNA we may expect them to match perfectly (or nearly so) but, as we will see,
there are often discrepancies.
■ Using Sanger sequencing, DNA is usually sequenced on both strands. However,
ESTs are often sequenced on one strand only, and therefore have a high error rate.
52

GenInfo Number
■ In addition to accession numbers, NCBI also assigns unique sequence
identification numbers that apply to the individual sequences within a record.
■ GenInfo (GI) numbers are assigned consecutively to each sequence that is
processed. For example, the human beta globin DNA sequence associated with the
accession number NM_000518.4 has a gene identifier GI:28302128. The suffix .4
on the accession number refers to a version number; NM_000518.3 has a
different gene identifier, GI: 13788565.
53

Tips for using ENTREZ database; (read)
■ The Boolean operators AND, OR, and NOT must be capitalized. By default, AND is
assumed to connect two terms; subject terms are automatically combined.
■ Perform a search of a specific phrase by adding quotation marks. This may
potentially restrict the output, so it is a good idea to repeat a search with and
without quotation marks.
■ Boolean operators are processed from left to right. If you add parentheses, the
enclosed terms will be processed as a unit rather than sequentially. A search of
NCBI Gene with the query “globin AND promoter OR enhancer” yields 31,000
results; however, by adding parantheses, the query “globin AND (promoter OR
enhancer)” yields just 66 results.
■ If interested in obtaining results from a particular organism (or from any taxonomic
group such as the primates or viruses), try beginning with TaxBrowser to select the
organism first. Adding the search term human[ORGN] will restrict the output to
human. Alternatively, you can use the taxonomy identifier for human, 9606:
txid9606[Organism:exp]
54

■ A variety of limiters can be added. In NCBI Protein, the search
500000:999999[Molecular weight] will return proteins having a molecular weight
from 500,000 to 1 million daltons. To view proteins between 10,000 and 50,000
daltons that I have worked on, enter 010000:050000[Molecular weight] pevsner j
(or, equivalently, 010000[MOLWT] : 050000[MOLWT] AND pevsner j[Author]).
■ By truncating a query with an asterisk, you can search for all records that begin
with a particular text string. For example, a search of NCBI Nucleotide with the
query “globin” returns 6777 results; querying with “glob*” returns 490,358 results.
These include entries with the species Chaetomium globosum or the word global.
■ Keep in mind that any Entrez query can be applied to a BLAST search to restrict its
output.
55

The Reference Sequence (refSeq) Project:
■ One of the most important developments in the management of molecular sequences
is RefSeq. The goal of RefSeq is to provide the best representative sequence for each
normal (i.e., nonmutated) transcript produced by a gene and for each normal protein
product.
■ There may be hundreds of GenBank accession numbers corresponding to a gene, since
GenBank is an archival database that is often highly redundant. However, there will be
only one RefSeq entry corresponding to a given gene or gene product, or several
RefSeq entries if there are splice variants or distinct loci.
■ Consider human myoglobin as an example. There are three RefSeq entries
(NM_005368.2, NM_203377.1, and NM_203378.1), each corresponding to a distinct
splice variant. Each splice variant involves the transcription of different exons from a
single-gene locus.
■ In this example, all three transcripts happen to encode an identical protein having the
same amino acid sequence. The source of the transcript distinctly varies, and may be
regulated and expressed under different physiological conditions. It therefore makes
sense that each protein sequence, although having an identical string of amino acid
residues, is assigned its own protein accession number (NP_005359.1, NP_976311.1,
and NP_976312.1, respectively).
56

■ RefSeq entries are curated by the staff at NCBI and are nearly nonredundant.
RefSeq entries have different status levels (predicted, provisional, and reviewed),
but in each case the RefSeq entry is intended to unify the sequence records.
■ You can recognize a RefSeq accession by its format, such as NP_000509 (P stands
for beta globin protein) or NM_006744 (for beta globin mRNA). The corresponding
XP_12345 and XM_12345 formats imply that the sequences are not based on
experimental evidence. A variety of RefSeq formats are shown in table 2.7 and
identifiers corresponding to human beta globin are shown in table 2.8.
■ A GenBank or RefSeq accession number refers to the most recent version of a
given sequence. For example, NM_000558.3 is currently a RefSeq identifier for
human alpha 1 hemoglobin.
■ We mentioned above that a suffix such as “.3” is the version number. By default, if
you do not specify a version number then the most recent version is provided.
alpha 1 hemoglobin. We mentioned above that a suffix such as “.3” is the version
number. By default, if you do not specify a version number then the most recent
version is provided.
57

RefSeq Gene and the Locus Reference Genomic
(LRG) project:
■ While the RefSeq project has a critical role in defining reference sequences, it has
several limitations. The changing version numbers of some sequences can lead to
ambiguity when scientists report RefSeq accession numbers without their version
numbers.
■ For example, a patient may have a variant at a specific nucleotide position in the
beta globin gene corresponding to NM_000518.3 but the version number is not
given. Once the record is subsequently updated to NM_000518.4, anyone studying
this variant might be unsure of the correct position of the variant since it depends
on which sequence version was used. To address these concerns about gene
variant reporting, the Locus Reference Genomic (LRG) sequence format was
introduced.
■ The goal of this project is to define genomic sequences that can be used as
reference standards for genes, representing a standard allele. No version numbers
are used and sequence records are stable and designed to be independent of
updates to reference genome assemblies. In a related response to this issue, the
RefSeq project was expanded to include RefSeqGene. 58

The Consensus Coding Sequence CCDS
project:
■ The Consensus Coding Sequence (CCDS) project was established to identify a core
set of protein coding sequences that provide a basis for a standard set of gene
annotations.
■ The CCDS project is a collaboration between four groups (EBI, NCBI, the Wellcome
Trust Sanger Institute and the University of California, Santa Cruz or UCSC).
■ Currently, the CCDS project has been applied to the human and mouse genomes;
its scope is considerably more limited than RefSeq. Its strength is that it offers a
“gold standard” of best supported gene and protein annotations with extensive
manual annotation by experts, enhancing the quality of the database.
59

The Vertebrate Genome Annotation (VEGA)
project:
■ It is essential to correctly annotate each genome; in particular, we need to define
gene loci and all their features. The Vertebrate Genome Annotation (VEGA)
database offers high-quality, manual (expert) annotation of the human and mouse
genomes, as well as selected other vertebrate genomes.
■ Performing a search for HBB at the VEGA website, there is one human entry. This
includes two main displays:
(1) a transcript view which provides information such as cDNA and coding sequences
and protein domain information; and
(2) a gene view which includes data on orthologs and alternative alleles.
60

Access to Information VIA Gene Resource at
NCBI:
■ How can one navigate through the bewildering number of protein and DNA
sequences in the various databases?
■ An emerging feature is that databases are increasingly interconnected, providing a
variety of convenient links to each other and to algorithms that are useful for DNA,
RNA, and protein analysis.
■ NCBI’s Gene resource (formerly called Entrez Gene, and LocusLink before that) is
particularly useful as a major portal. It is a curated database containing descriptive
information about genetic loci.
■ You can obtain information on official nomenclature, aliases, sequence
accessions, phenotypes, Enzyme Commission (EC) numbers, OMIM numbers,
UniGene clusters, HomoloGene (a database that reports eukaryotic orthologs),
map locations, and related websites.
61

■ To illustrate the use of NCBI Gene we search for human beta globin. The result of
entering an NCBI Gene search is shown in Figure 2.8. Note that in performing this
search, it can be convenient to restrict the search to a particular organism of
interest. (This can be done using the “limits” tab on the NCBI Gene page.) The
“Links” button (Fig. 2.8, top right) provides access to various other databases
entries on beta globin.
62

Clicking on the main link to the human beta globin entry results in the following
information (Fig. 2.9):
■ At the top right, there is a table of contents for the NCBI Gene beta globin entry.
Below it are further links to beta globin entries in NCBI databases (e.g., protein and
nucleotide databases and PubMed), as well as external databases (e.g., Ensembl
and UCSC).
■ Gene provides the official symbol (HBB) and name for human beta globin.
■ A schematic overview of the gene structure is provided, hyperlinked to the Map
Viewer (see “The Map Viewer at NCBI” below).
■ There is a brief description of the function of beta globin, defining it as a carrier
protein of the globin family.
■ The Reference Sequence (RefSeq) and GenBank accession numbers are provided
64

■ Figure 2.10 shows the standard, default form of a typical NCBI Protein record (for
beta globin).
■ It is simple to obtain a variety of formats by changing the display options. By
clicking a tab (Fig. 2.10a) the commonly used FASTA format for protein (or DNA)
sequences can be obtained, as shown in Figure 2.11.
■ Note also that by clicking the CDS (coding sequence) link of an NCBI Protein or
NCBI Nucleotide record (shown in Fig. 2.10b at the upper left), the nucleotides that
encode a particular protein, typically beginning with a start methionine (ATG) and
ending with a stop codon (TAG, TAA, or TGA), can be obtained.
■ This can be useful for a variety of applications including multiple sequence
alignment and molecular phylogeny.
66

Relationship between NCBI Gene, Nucleotide,
and Protein Resources:
■ If interested in obtaining information about a particular DNA or protein sequence, it
is reasonable to visit NCBI Nucleotide or NCBI Protein and perform a search.
■ A variety of search strategies are available, such as limiting the output to a
particular organism or taxonomic group of interest, or limiting the output to RefSeq
entries.
■ There are also many advantages to beginning your search through NCBI Gene. The
official gene name can be identified there, and you can be assured of the
chromosomal location of the gene.
■ Furthermore, each Gene entry includes a section of reference sequences that
provides all the DNA and protein variants that are assigned RefSeq accession
numbers.
69

Comparison of NCBI’s Gene and UniGene
■ As described above, the UniGene project assigns one cluster of sequences to one
gene.
■ For example, for HBB there is one UniGene entry with the UniGene accession
number Hs.523443. This UniGene entry includes a list of all the GenBank entries,
including ESTs, that correspond to the HBB gene.
■ The UniGene entry also includes mapping information, homologies, and expression
information (i.e., a list of the tissues from which cDNA libraries were generated
that contain ESTs corresponding to the RBP gene).
70

■ UniGene and NCBI Gene have features in common, such as links to OMIM,
homologs, and mapping information. They both show RefSeq accession numbers.
■ There are four main differences between UniGene and NCBI Gene:
1. UniGene has detailed expression information; the regional distributions of cDNA
libraries from which particular ESTs have been sequenced are listed.
2. UniGene lists ESTs corresponding to a gene, allowing them to be studied in detail.
3. Gene may provide a more stable description of a particular gene; as described
above, UniGene entries may be collapsed as genome-sequencing efforts proceed.
4. Gene has fewer entries than UniGene, but these entries are more richly curated.
NCBI Gene now has >200,000 human entries (as of 2015). These include gene
predictions, pseudogenes, and mapped phenotypes.
71

NCBI’s gene & Homologene:
■ The HomoloGene database provides groups of annotated proteins from a set of
completely sequenced eukaryotic genomes.
■ Proteins are compared placed in groups of homologs, and the protein alignments
are then matched to the corresponding DNA sequences.
■ You can find a HomoloGene entry for a gene/protein of interest by following a link
on the NCBI Gene page. A search of HomoloGene with the term hemoglobin results
in dozens of matches for myoglobin, alpha globin, and beta globin.
■ By clicking on the beta globin group, access can be gained to a list of proteins with
RefSeq accession numbers from human, chimpanzee, dog, mouse, and chicken.
■ The pairwise alignment scores are summarized and linked, and the sequences can
be downloaded (in genomic DNA, mRNA, and protein formats) and displayed as a
protein multiple sequence alignment.
72

Biological Databases | Access to sequence data and related information

More Related Content

What's hot (20)

Similar to Biological Databases | Access to sequence data and related information (20)

Recently uploaded (20)

Biological Databases | Access to sequence data and related information