Databases and Systems

There are other sets of data that can be grouped together. For example, sequences
of the same gene from a group of related organisms comprise a population set, such
as the actin sequences from related vertebrates. These and other more complicated
sets are just beginning to be used and will not be discussed further in this chapter.
Sequence Identifiers.
There are a variety of reasons why there are multiple sequence identifiers in the “id”
blocks of Bioseqs. For example, consider an expressed sequence tag (EST)
sequence. In its “native” database, dbEST[10], it is given an integral tag, an “est-id”
for its unique sequence identifier. If this sequence is to be used outside of the context
of this particular database, the string “dbEST” must be added to this integral
identifier for the resulting composite identifier to be unique. However, to appear
in GenBank[11], the sequence record needs, additionally, both a LOCUS name and
an ACCESSION number. Every sequence in GenBank is also given a “gi” number,
which allows all sequences to be retrieved with a single integral key, by Entrez, for
example. Data from other sources also retain their identifiers assigned by those
sources. So this single sequence record will have four to six different sequence
identifiers. Different retrieval programs will use a different identifier from this
synonymous set.
‘gi’s Vs accessions
There has been some confusion about the role of the ‘gi’ and how it compliments an
ACCESSION. When a laboratory submits a sequence from a piece of DNA to a
database, an ACCESSION number is assigned and is permanently associated with
that piece of DNA. When the record containing that sequence is loaded into the ID
database (see below) an initial ‘gi’ is assigned to that sequence. Further experiments
over a time may alter the best understanding of the true sequence of that DNA.
When these new sequences for the same piece of DNA are submitted to NCBI, a new
‘gi’ is assigned. This leads to a “chain” or series of sequences and corresponding
‘gi’s for the same piece of DNA. When it is important to identify the piece of DNA,
for example as a subclone at a particular location within some other clone, then the
ACCESSION is best used. When the particular sequence is most important, for
statements about sequence similarity or some conceptual translation, the ‘gi’ that
points to the particular sequence intended is best used. Statements and experiments
that use ‘gi’s can always be repeated, because the sequence identified with a
particular ‘gi’ is always available, even if the particular sequence identified by that
‘gi’ is not thought, currently, to be the accurate sequence. This relationship between
ACCESSION and ‘gi’ is shown in Figure 2, below.
gi gi gi
100 citation change 100 sequence change 201
Figure 2: A “chain” of ‘gi’s
15