Physical maps built from large-insert BAC clones and anchored to
linkage maps assist sequence assembly in whole genome shotgun (WGS)
sequencing projects , enable positional cloning of genes/QTLs and structural studies on gene families [2,3], and facilitate the isolation of homologous genes from model plants in heterozygous or polyploid species .
In plants, physical maps have been constructed for Arabidopsis thaliana , sorghum , rice , soybean , apple , black cottonwood , and grapevine .
Strategies based on BAC fingerprints detect overlaps among BAC clones
for the development of physical maps. Briefly, BAC clones are digested
with restriction enzymes and the fragments are separated by
electrophoresis, producing a pattern of bands. The overlap between
adjacent clones is identified by pairwise comparison of band profiles
and calculation of the proportion of shared bands .
Technologies for producing BAC fingerprints have evolved rapidly, from
using one to five restriction enzymes, and moving from agarose-based
gel to sequencer-based electrophoresis to generate the profiles .
The methods that combine the use of several restriction enzymes and
acrylamide gels or sequencers are usually referred to as High
Information Content Fingerprinting (HICF), because they have allowed a
dramatic increase in the sensitivity of the process. Since all of these
methodologies were not applied to the same biological materials, the
performances are not homogeneously comparable, and the debate on the
advantages and disadvantages of the different protocols still persists . For instance, it appears that the five-enzyme restriction protocol developed by 
leads to a higher error rate per fingerprint, but provides the highest
sensitivity compared with alternative techniques based on two or
three-enzyme restrictions . In the present work, the five-enzyme restriction protocol was adopted because of its higher sensitivity and throughput.
The grapevine genome has recently disclosed two peculiarities:
grapes are highly heterozygous and they have descended from an ancient
hexaploid ancestor [1,16,17]. The polyploid origin of the grapevine genome was revealed by whole proteome comparison 
but was undetectable using STS markers or nucleotide alignments. Thus
homeologous regions are not expected to hamper the construction of a
physical map of the grapevine genome, as their respective fingerprints
are substantially different. In turn, heterozygosity is likely to
affect the correct assembly of BAC fingerprints, as it did in the DNA
assembly of Ciona savignyi . This aspect was addressed in poplar by , but was somewhat neglected in grapevine  and in apple .
Here, a thorough analysis of the effects of heterozygosity on physical
map construction is presented that unveils contig features that were
not previously described in poplar. It is also shown that the final map
is an effective tool for mapping candidate genes for agronomic traits
like disease resistance, as well as for developing new genetic markers.
Strengthening the resistance to diseases is one of the major objectives in grapevine breeding .
Two types of defence can be categorized in plants, based on the width
of the host range. Non-host resistance is effective across an entire
plant taxon against all isolates of a pathogen. Host
resistance, the second type of resistance, is exerted at a
genotype-to-genotype level: only some of the genotypes of a plant taxon to
which a pathogen has adapted are resistant to any or all pathogenic
strains. This classification agrees well with that based on the type of
mechanisms and genes involved. Pre-invasion barriers and reactions
triggered by pathogen-associated molecular patterns (PAMPs), called
PAMP-triggered immunity (PTI), disrupt the potential ability of a
pathogen to attack a plant. Overall, the concepts of non-host
resistance and PTI overlap. Host resistance is a second line of defence
towards pathogens that have gained the capability of suppressing basal
or PAMP-triggered resistance. Host resistance can be either complete or
partial. The mechanisms underlying complete resistance are similar
across many types of plants. Complete host resistance is conferred by R
proteins that recognise pathogen effectors/suppressors or modifications
of their cellular targets (effector triggered immunity, ETI). R genes
are mostly arrayed in clusters, a physical organisation that generates
new variants at a rate higher than in any other class of genes [20,21]. The links between ETI, PTI, complete host, and non-host resistance were modelled by . Both PTI and ETI rely first on pathogen recognition carried out by receptors, consisting of transmembrane proteins for PTI [23,24],
and cytoplasmic proteins with a nucleotide-binding site and a
leucine-rich repeat domain (NBS-LRR) or receptor-like kinases (RLK) for
ETI . These two sides of the immune system are connected by proteins of the downstream signalling pathways, such as SGT1 and RAR1 [26-28], and by gene products of the salicilic acid (SA), jasmonic acid (JA), ethylene (ET), and MAPK cascade pathways .
Non-host resistance is the outcome of a heterogeneous set of genes,
and the proteins they specify, that are implicated in pathogen
accessibility/inaccessibility (i.e. lipase-like EDS1, synthaxin-like PEN1, etc.), cytoskeletal rearrangements and protein turnover (i.e. SGT1), PAMP-triggered responses, and synthesis of toxic metabolites [30-32].
The location of their homologues in the grapevine genome was recently
established using information from the grapevine genome sequence [1,17],
which was unknown when this study began. In contrast, the genes
triggering complete host resistance all conform to a few classes of
receptor-coding genes that are functionally similar. Clues on the size
and the genomic organisation of R gene families in grape were given by
genetic map data [2,33,34] and by a survey in the draft genome sequence .
In this paper, we present (1) the assembly of a 'Cabernet Sauvignon'
physical map based on restriction enzyme BAC fingerprinting, (2) the
anchorage of the physical contigs on the meiotic linkage maps, (3) the
use of this map for the placement of candidate genes for disease
resistance, and (4) the alignment between the physical location of
resistance gene analogues and phenotypic loci for pathogen
resistance based on bridging markers. How the high heterozygosity in
the 'Cabernet Sauvignon' genome has shaped some features of the
physical map is also studied in detail and discussed.