Datensatz

Zusammenfassung

Site-specific recombinases have become a resourceful tool for genome engineering, allowing sophisticated in vivo DNA modifications and rearrangements, including the precise removal of integrated retroviruses from host genomes. In a recent study, a mutant form of Cre recombinase has been used to excise the provirus of a specific HIV-1 strain from the human genome. To achieve provirus excision, the Cre recombinase had to be evolved to recombine an asymmetric locus of recombination (lox)-like sequence present in the long terminal repeat (LTR) regions of a HIV-1 strain. One pre-requisite for this type of work is the identification of degenerate lox-like sites in genomic sequences. Given their nature-two inverted repeats flanking a spacer of variable length-existing search tools like BLAST or RepeatMasker perform poorly. To address this lack of available algorithms, we have developed the web-server SeLOX, which can identify degenerate lox-like sites within genomic sequences. SeLOX calculates a position weight matrix based on lox-like sequences, which is used to search genomic sequences. For computational efficiency, we transform sequences into binary space, which allows us to use a bit-wise AND Boolean operator for comparisons. Next to finding lox-like sites for Cre type recombinases in HIV LTR sequences, we have used SeLOX to identify lox-like sites in HIV LTRs for six yeast recombinases. We finally demonstrate the general usefulness of SeLOX in identifying lox-like sequences in large genomes by searching Cre type recombination sites in the entire human genome. SeLOX is freely available at http://selox.mpi-cbg.de/cgi-bin/selox/index.