Genomic Libraries

 Early genomic libraries were viewed as a repository from which any genetic elements could be isolated for a given organism. Modern genomic libraries serve this function, but they also straddle the interface between genomic bioinformatics and functional biology. Thus, genomic sequences are increasingly understood in terms of their information content, comprising their RNA- and protein-coding sequences, as well as in terms of the overall organization of their embedded genetic program. Central to this information-based approach are new, highly sophisticated libraries of genomic DNA that maintain the structural context of genetic elements over several hundred kilobase pairs or more. Such libraries are essential for gene mapping projects, which will eventually produce an atlas of the genetic organization of many organisms. Detailed gene maps are required to interpret sequencing data that result from genome sequencing efforts. The genome sequencing projects themselves depend on high-quality genomic DNA libraries as a source of genetic material. The study of the genetic basis of disease makes use of sequence differences between individuals that can be ascertained through genomic libraries. These libraries also serve as a source of genetic probes that can be used for diagnosis and characterization of the relevant genotype of any individual.

 Construction of genomic libraries requires particular attention during all stages of processing in order to minimize mechanical shearing of the cellular DNA. Nucleic acids become increasingly sensitive to shearing with increasing size, making it especially difficult to maintain the integrity of chromosome-sized species. Fortunately, specialized methods have been developed that enable the cloning of DNA fragments that are several hundred kilobase pairs or larger. The size of the genomic insert determines the type of vector into which the DNA is inserted. Inserts smaller than 10 kbp can be cloned into either plasmid or bacteriophage vectors. Cosmid vectors are required for fragments in the range of 10 to 40 kbp, while artificial chromosomes (ACs) are needed for even larger inserts.

 Most genomic DNA library projects seek to represent a set of partially overlapping DNA fragments that collectively represent the entire genome of the subject organism. Such overlapping cloned inserts are ideal for chromosome “walking” approaches, in which the termini of one set of probes can be used to identify overlapping clones, which in turn can identify further overlapping clones downstream. In this manner, a set of clones can be derived that cover virtually any span of genetic sequences. Once prepared, such a repository facilitates the identification of genetic elements associated with any phenotype or disease. Three methods are employed for generating genomic DNA fragments of appropriate sizes for cloning. First, mechanical shearing can be used to produce relatively small inserts, although ligation into the vector with such fragments is typically inefficient. Because the shearing is random, this method produces a series of partially overlapping fragments. Second, and by far the most common method, is the use of restriction enzymes to prepare fragments. Digestion of DNA with restriction enzymes provides a great deal of control over the size of the resulting fragments. For example, endonucleases with a four-nucleotide recognition sequence cut on average every 256 nucleotides, while certain enzymes with large recognition sequences may cut approximately every 10⁶ nucleotides. Unfortunately, complete digestion of genomic DNA with restriction enzymes eliminates any overlapping fragments; therefore, most genomic libraries employ partial digestion of genomic DNA. The degree of digestion thus influences the average fragment size, as well as the degree of fragment overlap. Another advantage of using restriction endonucleases is that the fragment termini are defined and readily clonable. The third method of preparing genomic DNA is to perform PCR using primers containing a region of random sequence. In the first PCR cycle, the random primer segments anneal to both strands throughout the genome and are extended in a 5′ to 3′ direction. In subsequent PCR cycles, any two oppositely oriented primers that are within sufficient proximity can generate a specific PCR product. The resulting PCR products can be prepared for cloning into an appropriate vector. Alternatively, the collection of genomic DNA fragments can be maintained and propagated as PCR products using fixed sequences, introduced during the PCR, as primer binding sites for subsequent rounds of amplification.

In general, genomic libraries are used for DNA-based studies, such as gene mapping and marker analysis; however, in certain cases they may also be useful as expression libraries. Many simpler organisms have few or no introns, so their genomic DNA can be directly expressed into messenger RNA and translated into protein. Furthermore, structural RNAs, such as ribosomal RNA and transfer RNA are not represented in most cDNA libraries and thus require genomic libraries to isolate appropriate DNA coding sequences. Finally, random genomic expression can be a productive means of generating RNA and peptide libraries for epitope- and domain-mapping exercises.