"Frontmatter". In: Plant Genomics and Proteomics

I
NFORMATICS
T
OOLS
Once the data have been generated and appropriately stored in a database,
it is usually necessary to analyze that data in some form. A large number of
1 7 0
9. B
I O I N F O R M AT I C S

---

tools have been, and are continuing to be, developed for this purpose to
cover the whole range of genomics data that are acquired. Many of these
tools are available free over the Internet, and many of the sites that house
the data are also designed to allow access to the various tools so that the
analyses can all be performed through that particular site. In addition, many
of the tools are designed to give graphical representations/visualizations for
ease of display and interpretation of the results. 
I
NFORMATICS AND
A
NALYSES OF
P
LANT
G
ENOMES
Genomic approaches to plant biology have provided a tremendous wealth
of gene and gene sequence data for a wide variety of plant species. The avail-
ability of these large bodies of data provides both opportunities and chal-
lenges. The cross-referencing of the sequence data with the expression data
available from functional genomics and proteomics projects provides an
opportunity to develop a complete catalog of gene function across the plant
kingdom. A start has been made on this with the 2010 program at the NSF
and its counterparts in Europe and Japan. It is important to integrate diverse
data and data types to provide a more complete and consistent view of all
this information. Therefore, the annotation of the data must be consistent
and identifiable. However, the current crop of bioinformatics software does
not generate a single unambiguous conclusion, because they use different
criteria either for identifying the features associated with a particular
genomic sequence, for example, the identity of splice sites, or for recogniz-
ing the component open reading frames in a spliced gene. 
S
EQUENCE
A
SSEMBLY
Sequence assembly is a crucial early step in many genomics projects. This is
because most of the initial sequence “reads” are very much shorter than the
functional or important region of the genome being characterized. As
described in Chapter 3, whole genome sequencing projects use two differ-
ent approaches to generate the sequence, whole genome shotgun sequenc-
ing and minimum tiling path methods. However, in each case the sequence
reads must be assembled into large contiguous regions of genomic sequence.
The draft genomic sequences then must be finished. Finishing is the process
of turning a rough draft assembly into a highly accurate contiguous DNA
sequence with a defined maximum error rate. Therefore, finishing involves
closing the remaining gaps, resolving ambiguities, and validating the assem-
bly. The aim is that the overall sequence quality conforms to the Bermuda
standard of being confirmed by at least two templates, accurate to at least 
1 bp in 10,000, and with no gaps. Finishing usually makes use of multiple
software tools in an iterative manner to obtain contiguous sequence that
meets the standards for double-strand coverage and sequence quality
I
N F O R M AT I C S
T
O O L S
1 7 1

---

(http://www.genome.gov/10001812). Once the finished assembly is com-
plete it can be experimentally verified via PCR using selected primer pairs.
These primers are often the same ones that were generated to close any gaps
in the original assembly. The process of finishing a draft assembly is shown
in Figure 9.1. 
EST A
SSEMBLY
Sequence assembly is also used to collapse the large libraries of expressed
sequence tags (ESTs) to produce tentative contigs (TCs) or unigene sets
(TUGs). Such assembly procedures eliminate the redundancy that exists in
1 7 2
9. B
I O I N F O R M AT I C S
IMPORT DATA
DATA STORAGE
AND ASSEMBLY
RESEQUENCING
LOW QUALITY
REGIONS
READ INTO GAPS
GAP CLOSURE
Contig 1
Contig 2
Gap
Yes
No
Are clones available that cover gap?
Sequence these
Submit to GenBank
PCR across gap
Add to data storage
and assembly
Sequence from PCR product
or from subclones
Data Review
and quality
control
FIGURE 9.1.
Genome sequence finishing. The rough draft assembly is converted
into a fully assembled contiguous sequence by resequencing and gap closure
(Adapted from http://www-shgc.stanford.edu/Seq/doepages/methodology.html).

---

the EST data because multiple ESTs may possibly represent different parts
of the same gene. The clustering also can extend the length of the sequence
and would ideally result in the full length of the gene being included in the
final assembled product. Such collapsing of the data sets is important. Thus
the 415,000 wheat ESTs in GenBank are unlikely to represent an equivalent
number of separate genes given our knowledge of the gene content of other
plant genomes. 
The unigene set that is derived from the assembly of ESTs is not a
compete set of genes for that organism because the cDNA libraries used to
generate the ESTs will not reflect all the genes in a particular genome. The
expression of some genes is too low or too transient to result in capture,
whereas other genes will only be expressed in tissues under certain growth
conditions or in particular tissues. Therefore, the unigene set is simply the
least redundant set of expressed sequences that can be arrived at by using
all the available data. 
The extent of the coding regions included in the unigene set can be
extended by the use of gene predictions from complete genomic sequences.
The matching region can be aligned and gene prediction programs used to
identify the possible transcription and translation start sites. This informa-
tion can be used to design primers to be used in reverse transcription PCR
(RT-PCR) to test whether the predicted fragments can be amplified from
mRNA populations. Again, informatic analysis of the regions 5¢ to the ten-
tative expressed region may indicate where and/or when such a transcript
may be expressed, thereby narrowing the range of tissues or developmental
stages that must be screened to generate experimental evidence for the pre-
dicted expression. The 5¢ regions are the least represented transcribed
sequences, as cDNA, SAGE, or MPSS libraries are all biased toward the 
3¢ ends of the mRNAs because of the oligo-dT priming of the first-strand
synthesis.
T
HE
R
IGHT
T
OOL FOR THE
R
IGHT
J
OB
Many of the database sites contain a range of tools for sequence 
characterization and identification. The NCBI (http://www.
ncbi.nlm.nih.gov/BLAST/) site makes available a range of tools with a
regular update of all the documentation describing the programs, their uses,
and the interpretation of the resulting data. Many of the specific data-
bases also include BLAST as an integral option, although without the full
range of functions or the tutorials.
BLAST
BLAST
®
(Basic Local Alignment Search Tool) is a set of similarity search 
programs designed to explore all of the available sequence databases regard-
I
N F O R M AT I C S
T
O O L S
1 7 3

---

less of whether the query relates to protein or nucleotide sequence. 
The BLAST programs have been designed for speed, with a minimal 
sacrifice of sensitivity to distant sequence relationships. The scores assigned
in a BLAST search have a well-defined statistical interpretation, making real
matches easier to distinguish from random background hits
(http://www.ncbi.nlm.nih.gov/BLAST/blast_overview.html).
The BLAST algorithm does not only identify matches that serve to make
a correct functional assignment. BLAST has a large set of variable parame-
ters that can be altered to produce a range of different matches. The data-
base that is the target of the BLAST search changes constantly as the available
set of sequenced genes increases. This will also change the output, some-
times dramatically. Thus even a program as basic as BLAST will not gener-
ate the same results reproducibly unless users carefully coordinate and
synchronize all the program’s inputs. Although the NCBI site has a BLAST
Program Selection Guide (reproduced in Tables 9.1 and 9.2) and a series of
tutorials to aid in the use and interpretation of the data, the data still gen-
erally must be manually inspected to determine which particular matches
are meaningful. 
S
EQUENCE
M
ATCHING
Sequence similarity is a very general tool that forms the basis of many dif-
ferent biological sequence analyses. The basic tool for generating sequence
matches is BLAST. It is limited by the traditional alignment presentation
style of the results. An alternative program such as Miropeats
(http://www.genome.ou.edu/miropeats.html) discovers regions of
sequence similarity within or among any set of DNA sequences and then
graphically represents the regions that are similar. An example is given in
Figure 9.2, which shows the distribution of repeated segments over a short
stretch of flax sequence, with a pair of inverted repeats, a tandem oligo
repeat, and a palindrome. In Figure 9.3, the same region of the Arabidopsis
chromosome 2 (BAC F16P2) that is shown in Figure 1.4 is shown as a Miro-
peat pattern rather than with the repeats indicated with directed arrows as
in Figure 1.4. The enhancement offered by Miropeats when making con-
ventional DNA sequence comparisons is the summary of extensive large-
scale sequence similarities on a single page of graphics. Miropeats can
handle the comparison of the repeat structures of entire chromosomes, visu-
alizing overlapping sequence fragments in a contig assembly project or com-
paring the products of different contig assembly programs.
A
NNOTATIONS OF
G
ENOME
S
EQUENCE
Once the genome sequence has been assembled, the important features such
as genes, transposons, and repeats, must be placed on the sequence. This
1 7 4
9. B
I O I N F O R M AT I C S

---

I
N F O R M AT I C S
T
O O L S
1 7 5
TABLE 9.1. W
HICH
BLAST 
TOOL IS
A
PPROPRIATE FOR
S
PECIFIC
S
EARCHES
? 
If your sequence is NUCLEOTIDE
Length
Database
Purpose
BLAST Program
20 bp or 
Nucleotide
Identify the query
MEGABLAST (accept
longer
sequence
batch queries)
Standard BLAST
(blastn)
Find sequences similar
Standard BLAST
to query sequence
(blastn)
Find proteins similar to
Translated BLAST
translated query in a
(tblastx)
translated database
Protein 
Find proteins similar
Translated BLAST
to translated query
(blastx)
in a protein database
7–20 bp
Nucleotide
Find primer binding
Search for short,
sites or map short
nearly exact matches
contiguous motifs
If your sequence is PROTEIN
Length
Database
Purpose
BLAST program
15 residues
Protein
Identify the query 
Standard Protein BLAST
or longer
sequence or find
(blastp)
protein sequences
similar to query
Find members of a 
PSI-BLAST
protein family or build
a custom position-
specific score matrix
Find proteins similar to
PHI-BLAST 
the query around a
given pattern
Conserved
Find conserved domains
CD-search
Domains
in the query
(RPS-BLAST)
Conserved
Find conserved domains 
Conserved Domain 
Domains
in the query and
Architecture 
identify other proteins
Retrieval Tool
with similar domain
(CDART)
architectures
Nucleotide
Find similar proteins in 
Translated BLAST
a translated nucleotide
(tblastn)
database 
5–15 residues
Protein
Search for peptide motifs
Search for short, nearly
exact matches
From http://www.ncbi.nlm.nih.gov/BLAST/producttable.html

---

1 7 6
9. B
I O I N F O R M AT I C S
T
ABLE 9.2. S
PECIALIZED
D
A
T
AB
ASES
A
V
AILABLE
FROM
NCBI
Specialized Database Sear
ches
Query
Database
Purpose
BLAST Pr
ogram
Nucleotide 
None
Compar
e the query and second sequence
BLAST 2 Sequences 
or Pr
otein
dir
ectly
The NCBI Draft 
Map the query sequence. Determine the
Human Genome BLAST 
Human Genome
genomic str
uctur
e. Identify novel genes.
Mouse Genome
Map the query sequence. Determine the
Mouse Genome BLAST 
genomicstr
uctur
e. Identify novel genes.
Rat
Map the query sequence. Determine the
Rat Genome BLAST page 
genomic str
uctur
e. Identify novel genes.