Sistema de almacenamiento y transferencia de crudo

A critical step in positional cloning is the identification of candidate genes from a large, genetically defined region. The most commonly used methods include isolation o f CpG islands, exon trapping, cDNA selection, and genomic sequencing followed by database comparisons as well as exon prediction through computational analysis of the sequence.

The development and map assignment o f ESTs, which are being developed to identify expressed genes, is facilitating the construction of a gene map, thus simplifying gene identification.

1.6.1 Detection of CpG islands

CpG islands are short GC-rich sequences about 1 Kb long, which are unusually hypomethylated and are often found at 5’ ends of vertebrate genes (Bird 1987). They can be easily recognised by a variety o f rare-cutter restriction enzymes whose recognition site contains one or two CpG dinucleotides. This property makes them useful landmarks for identifying novel genes (Lindsay and Bird 1987). However, this method is not always valid since a substantial number o f genes are not associated with CpG islands.

1.6.2 Exon trapping

The exon trapping strategy to identify genes takes advantage o f the property o f exons to engage in an artificial RNA splicing assay (Duyk et al. 1990, Buckler et al. 1991).

Although exon trapping has been used in the identification o f many genes, including the Huntington’s disease gene (the Huntington’s Disease Collaborative Research Group, 1993) and the neurofibromatosis type 2 tumour suppressor gene (Trofatter et al. 1993), it is not always successful and is best used in combination with other strategies.

1.6.3 cDNA selection

cDNA selection involves the formation o f DNA/cDNA heteroduplexes, by hybridising a PCR-amplified cDNA library to a cloned genomic DNA, such as a YAC insert, immobilised on a filter (Lovett et al. 1991, Pamiroo et al. 1991).

This technique has the advantage of enabling detection o f rare or tissue specific transcripts and can be applied to several genomic clones simultaneously. A drawback is that since short fragment cDNAs constitute the selected material, to obtain a full-length coverage of a gene an entire cDNA library needs to be subsequently screened.

1.6.4 Computer-based DNA sequence analysis

Once the sequence of a DNA clone is known, computational analysis provides a powerful approach to determine whether that sequence is likely to be part o f a gene. The analysis can be performed with two main types of software. The first is based on homology programs such as BLAST (Altschul et al. 1990) and PASTA (Pearson and Lipman 1988) (section 2.8.2), which searches sequence databases for coding sequences that shares similarities with the given DNA sequence both between and within species. The second is based on exon-prediction programs such as GRAIL (Uberbacher and Mural 1991), and looks for sequences in the genomic clone that share similarities with gene-specific features (i.e. ORFs, CpG islands, promoter regions, intron-exon boundaries, poly-A sites). Generally, the two types o f computer programs are combined using integrated software packages such as NIX (section 2.8.4), to achieve maximum results.

Some o f the limitations of in silico predictions include accurate definition o f 5’ and 3’ ends of genes, and the identification o f intronless transcripts and genes with small ORFs.

1.6.5 Expressed sequence tags and full-length cDNAs

The identification o f a positional candidate in a disease-associated region is now often reduced to searching computer databases where lists o f genes mapping to specific chromosomal regions can be obtained. The development o f transcript maps, which detail the location o f transcribed sequences, incorporating all available expressed sequence tags (ESTs), has made this approach more feasible. ESTs are partial sequences generated from cloned cDNA, which are being developed to allow rapid identification o f expressed genes by sequence analysis. Alongside mapping o f ESTs using radiation mapping, expression studies are also being performed, so that an increasing number o f ESTs come with chromosomal location and expression pattern information. With the increasing number o f ESTs being produced, contributing to the transcript map o f the human genome, selection o f ESTs as positional candidates is taking over from the “traditional” gene isolation techniques described previously.

Over several years Adams et a l (1991, 1992, 1993, 1995) contributed massively to the generation o f ESTs with the production o f almost 200,000 sequences. Their general approach was to generate ESTs from random primed and partial cDNA clones rather than sequencing the ends of full-length cDNA clones, which contain 5’ and 3 ’ untranslated sequences. Other genome centres who contributed to the EST sequence databases include the Genexpress Centre (Genethon, Houlgatte et al. 1995), creating 18,698 ESTs, and the Washington-University-Merck EST project in collaboration with the IMAGE consortium (Integrated Molecular Analysis of Genomes and their Expression, Lennon et al. 1996), with more than 300,000 ESTs derived from both 5’ and 3’ ends o f cDNAs (Hillier et al. 1996). Most accumulated sequences are gathered in the dbEST division o f the Genbank database (Boguski et al. 1993), which now counts 3,169,953 human ESTs. Several genome centres have reported the assembling o f deposited ESTs and gene sequences into virtual transcripts representing, in some cases, full-length cDNAs. Full length cDNAs would prove more useful than the short EST sequences produced, which have often been associated with a variety of artefacts including genomic contamination, sequencing errors and improper splicing (Wolfsberg and Landsman 1997). Among the Genome centres involved in this effort is The Institute for Genomic Research (TIGR) and the UniGene division o f the National Centre for Biotechnology Information (NCBI). However, because expression of a single gene may

culminate in production o f several different mRNA transcripts, depending both on the gene and the source tissue, and considering also the technical challenges o f converting mRNAs into cDNAs, libraries with abundant truncated products are the common result, particularly for longer mRNAs. Different groups are therefore generating and sequencing sets o f unique full-length cDNA clones, which are having a major impact on the positional cloning o f disease-associated genes. The European IMAGE consortium (EURO-IMAGE, (http ://www.ornl.gov/meetings/wccs/euro.htm), founded in 1997, is among these groups. Their major goals include generating a set o f non redundant cDNA clones for most human gene transcripts with a master set o f unique full-length cDNA clones based upon the IMAGE Consortium resources, and obtaining high resolution and comparative functional mapping localisation in man with model organisms o f genes represented in the master set. The Kazusa DNA Research Institute (Nomura et al. 1994, Nagase et al. 2000) is also conducting a large- scale collection and sequencing of full-length cDNA by isolating clones from full-length enriched human cDNA libraries made by an "Oligo-capping" method. This method involves replacement o f the cap structure of mRNAs with an oligoribonucleotide. The oligo-capped mRNA is the used for first strand synthesis with dT adapter primers to construct full-length cDNA libraries. The 5’ end sequences of the clones are determined and used to select putatively full-length cDNA clones, which are then entirely sequenced. Their aim is to determine the sequence of 20,000 full-length cDNA clones. The German Genome Project also founded a cDNA consortium in 1997, with the aim of characterising the complete sequences o f novel human transcripts at the cDNA level. They have recently characterised 500 novel full-length cDNAs and over 1000 cDNAs corresponding to previously known genes (Wiemann et al. 2001), by using a method similar to the one employed by the Kazusa Institute. The full-length cDNA is a crucial tool both for the annotation of the human genome and for the experimental analysis o f gene function, as it enables the definition o f precise protein coding sequences and, in conjunction with the genomic sequence, to define gene structures and the composition o f exons in alternatively spliced transcripts o f the same gene. Once a full gene catalogue with full-length cDNA sequences covering every human gene has been established it will be possible to undertake large-scale and comprehensive functional analysis o f human genes and proteins.

The power of comparative genomics is already being realised with the sequencing of other genomes, such as Drosophila, mouse, zebrafish and chicken (Adams et al. 2000), along with genomes o f organisms already fully sequenced, i.e. Saccharomyces cerevisiae, Arabidopsis thaliana and Caenorhabditis elegans (Dujon 1996, the Arabidopsis Genome Initiative 2000, the C. elegans Sequencing consortium 1998). Parallel initiatives to characterise cDNA library ESTs from a variety of tissues for many species provides additional database resource for the characterisation of novel genes, particularly useful for understanding gene function. This allows high resolution and comparative functional mapping in human and model organisms o f several genes, and is often extremely useful in the isolation o f new human disease-genes.