Physical mapping of the human genome appears to be a complicated probl
em for the molecular genetics of higher organisms. The difficulties ar
e due not only to the great amount of work to be done, but mainly to t
he special features of genome organization (repeated nucleotide sequen
ces, nonclonable, methylated and harmful fragments, etc.). The drawbac
ks of molecular hybridization mapping arise from the repeats of differ
ent types common for the genome as a whole. Giant DNA fragments (e.g.,
YAC, Yeast Artificial Chromosomes) also cannot form the basis of geno
me mapping, because of their deletions, transformations, and chimerism
. A new strategy for genome mapping is described with reference to chr
omosome 3, one of the largest human chromosomes. Its main steps are th
e following: 1) sequencing of STS (Sequence-Tagged Sites) flanking DNA
rare-cutting restriction sites; 2) positioning of the STS along the c
hromosomal DNA (generation of the contigs) based on comparative comput
er analysis of STS from the same and from the neighboring restriction
sites, 3) determination of the distance between STS by hybridization w
ith large DNA fragments. A new family of vectors (SK series) was speci
ally designed for genome cloning and the simplified inexpensive techno
logy for preparing the jumping/linking libraries was developed within
this so-called shotgun sequencing strategy for extensive genome mappin
g. We have sequenced chromosome 3 fragments around NotI sites, and hav
e constructed the contigs covering about 50 Mbp of genomic DNA. Prospe
cts for mapping the total human genome are discussed.