"Frontmatter". In: Plant Genomics and Proteomics

G
ENERATING A
P
HYSICAL
M
AP
The current technology for generating a physical map usually involves
ordering a series of BAC clones (Marra et al., 1997). The BAC libraries can
be generated by using a number of different restriction enzymes (see Chapter
2) or by random shearing. The clones from these libraries are then finger-
printed. The fingerprinting involves isolating the BAC DNAs, digesting
them with a restriction enzyme, and running the fragments generated on a
gel. The overlap between different BACs is calculated from the number of
identically sized fragments that they have in common (Soderlund et al.,
2000). When a sufficient overlap between two BACs is found (the cutoff is
user defined), those two BACs are assumed to have the region containing
the identically sized bands in common and are therefore placed in a contig
(contiguous sequence) (Figure 3.2). The comparisons and assembly can be
automated with the use of image software to analyze the gels, and the fin-
G
E N E R AT I N G A
P
H Y S I C A L
M
A P
5 1

---

gerprint contigs (FPC) software is used to assemble the contigs (Soderlund
et al., 2000). All the data generated from thousands of BACs are assembled,
and the assembly constitutes the physical map. This assembly must be
checked by confirming that the genes, markers, or sequences that are known
to be on the BACs assembled into specific contigs are actually present in close
proximity on the genetic map. Ultimately, all of the data that have been accu-
mulated from the molecular mapping exercises can also be placed on the
physical map. Thus where a unique molecular marker is hybridized to the
BAC library, the selected fingerprinted BACs to which it hybridizes are
immediately anchored to the position of the chromosome assigned to the
marker. If the genetic and physical maps disagree, then the conflict must be
resolved to determine which assembly, the genetic map or the BAC contig,
is correct. Such physical maps based on fingerprinted contigs have been con-
structed for a number of species including humans, Arabidopsis, and rice. 
EST markers that have been mapped to BAC clones can be entered into
the FPC database. These data will help the assembly process, as well as
placing the ESTs on both anchored and unanchored contigs (Soderlund et
5 2
3. S
E Q U E N C I N G
S
T R AT E G I E S
1
2
3
4
5 6 7
8 9
10
11
12 13
14
15 16
17
a
b
c
d
e
f
g
h
i
k
l
m
n
o
p
q
r
FIGURE 3.2.
Illustration of the DNA fingerprinting method. Horizontal lines (a–r)
represent individual BAC clones that have been aligned based on shared restriction
fragment patterns. Vertical lines (1–17) represent the positions of restriction sites that
were used. For actual data, the comparison usually includes 20–40 bands. In the
figure the overlap between clones a and b is very likely to be correct. However,
without the intermediate clones b through h, the overlap between a and i would be
tenuous at best. As can be seen from the diagram, to be certain about the relative
overlaps the same region must be sampled multiple times, with 13 different BACs
contained within the length of a. 

---

al., 2000). The FPC software can use both the fingerprints and the markers
in generating the assembly, so if two clones share a marker, then a less strin-
gent overlap based on fingerprints will still be recorded as an overlap.
Because FPC can generate incremental updates, the contigs can be built as
the data are generated, rather than having to await all the data and having
a massive one-time final build of the physical map.
How many BACs are needed to achieve such a physical map (assuming
that there are no structural impediments to achieving the overlapping sets
such as identical duplicated regions)? If we assume that the BAC library was
generated with the average insert size of 125 kb, then a complete genome
would be contained in about 1000 BACs for Arabidopsis ( a genome size of
125 Mb), 3512 BACs for rice (genome size of 439 Mb), 21,728 BACs for maize
(genome size of 2,716 Mb), and 128,000 BACs for wheat (genome size of
16,000 Mb). As can be seen from Figure 3.2 multiple sets of the genome must
be fingerprinted, perhaps up to 20 times the number required for a complete
genome so that enough representatives from each region can be sampled.
Thus a twentyfold (20
x
) oversampling would mean that the numbers for Ara-
bidopsis rise to 20,000, for rice to 70,000, for maize to 435,000, and for wheat
to 2,560,000 BACs. Therefore, as the genome size increases the number of
BACs that must be fingerprinted to get some meaningful assembly of the
genome also rises. Even with this level of oversampling, most of the plant
genomes would not be assembled into the number of contigs that is the same
as the number of chromosomes (the ideal result). In general, the number of
contigs will be much larger than the chromosome number and the average
size of the contigs much smaller than the length of the chromosome. 
These contigs can then be placed on the genetic map by using molecu-
lar markers that hybridize to the BACs within a contig to determine the order
of the contigs along the chromosome. Obviously, for this ordering along the
genetic map to be successful, the spacing of the molecular markers has to be
less than the size of the contigs so that at least one marker is present on each
of the contigs. Alternatively, the BACs can be directly mapped onto the chro-
mosomes with fluorescent in situ hybridization (FISH).