268
LUIZ F. ONUCHIC AND GREGORY G. GERMINO
three times greater than that required by the vector. The highest cloning
yields are obtained when the partial-digestion products are size fractionated
prior to ligation to the vector. This step, however, can result in significant
losses of DNA if the method of size fractionation is not very efficient. Excessive manipulation of the DNA can further reduce cloning efficiency by
shearing molecules or producing ragged ends that are incompatible with the
vector. Lastly, inhibitors ofligation can be introduced if the reagents arenot
carefully prepared and purified. Sucrose gradient purification and preparative gel electrophoresis are the most commonly used fractionation methods.
Preparative gel electrophoresis offers better resolution and is capable of resolving larger quantities ofDNA. The sucrose gradient method, however, is
very simple, reliable and fast and is the method most commonly used in
genomic library construction. The basis for fractionation isthat DNA fragments migrate through a linear sucrose gradient at a rate dependent on their
size. This technique can resolve fragments in the range of 5 - 60 kb, which
are suitable for generating cosmid and bacteriophage libraries. Such a procedure is also used for purification oflambda bacteriophage arms. The steps
are decribed in the Procedure.
Randomly sheared libraries
Although partial-digest libraries generally lack fewer sequences than complete-digest libraries, there often are some cloning gaps. Figure 2 illustrates
one reason why this occurs. Restrietion sites are variably sensitive to cleavage. Some sites that always cut may be dustered together, always yielding
very small, unclonable fragments. Other sites may be very resistant and
yield fragments that are too large to clone under conditions that yield
an optimal size for the majority of fragments. Partial-digest libraries
lack fragments that are bounded by restriction sites that are either "hypersensitive" or "hyper-resistant" to the enzyme. There are several solutions to
this problem. One approach is to clone the DNA into vectors with different
insert size limits, eg. bacteriophage and cosmid. Another solution is to make
several libraries, using different enzymes for each. A final approach is to
avoid the use of enzymes altogether and resort to random shearing of
the source DNA. This process is thought tobe less biased in its inclusion
of fragments. Although this premise is unlikely to be completely true, it
undoubtedly is more random than restriction digestion. The major drawback to this approach is that sheared DNA has either ragged or blunt ends,
which require further manipulations that result in less efficient cloning
(Meyerowitz et al. 1981).
LUIZ F. ONUCHIC AND GREGORY G. GERMINO
three times greater than that required by the vector. The highest cloning
yields are obtained when the partial-digestion products are size fractionated
prior to ligation to the vector. This step, however, can result in significant
losses of DNA if the method of size fractionation is not very efficient. Excessive manipulation of the DNA can further reduce cloning efficiency by
shearing molecules or producing ragged ends that are incompatible with the
vector. Lastly, inhibitors ofligation can be introduced if the reagents arenot
carefully prepared and purified. Sucrose gradient purification and preparative gel electrophoresis are the most commonly used fractionation methods.
Preparative gel electrophoresis offers better resolution and is capable of resolving larger quantities ofDNA. The sucrose gradient method, however, is
very simple, reliable and fast and is the method most commonly used in
genomic library construction. The basis for fractionation isthat DNA fragments migrate through a linear sucrose gradient at a rate dependent on their
size. This technique can resolve fragments in the range of 5 - 60 kb, which
are suitable for generating cosmid and bacteriophage libraries. Such a procedure is also used for purification oflambda bacteriophage arms. The steps
are decribed in the Procedure.
Randomly sheared libraries
Although partial-digest libraries generally lack fewer sequences than complete-digest libraries, there often are some cloning gaps. Figure 2 illustrates
one reason why this occurs. Restrietion sites are variably sensitive to cleavage. Some sites that always cut may be dustered together, always yielding
very small, unclonable fragments. Other sites may be very resistant and
yield fragments that are too large to clone under conditions that yield
an optimal size for the majority of fragments. Partial-digest libraries
lack fragments that are bounded by restriction sites that are either "hypersensitive" or "hyper-resistant" to the enzyme. There are several solutions to
this problem. One approach is to clone the DNA into vectors with different
insert size limits, eg. bacteriophage and cosmid. Another solution is to make
several libraries, using different enzymes for each. A final approach is to
avoid the use of enzymes altogether and resort to random shearing of
the source DNA. This process is thought tobe less biased in its inclusion
of fragments. Although this premise is unlikely to be completely true, it
undoubtedly is more random than restriction digestion. The major drawback to this approach is that sheared DNA has either ragged or blunt ends,
which require further manipulations that result in less efficient cloning
(Meyerowitz et al. 1981).