next up previous
Next: Pruning Up: Operators Previous: Operators


Gene duplication involves making a copy of a gene or genes [Elseth 95], and has been used many times in Evolutionary computation [Schutz 97]. Duplicated genes can lie adjacent to the original gene, or can lie at a completely different region. Repeated gene duplications are presumed to have given rise to the human hemoglobin gene family. The hemoglobin protein is involved with transporting oxygen around the body through the bloodstream. The original hemoglobin gene duplicated to form an $\alpha$ - globin, and $\beta$ - globin genes. Each of these underwent further duplications resulting at present in four variations on the $\alpha$ - globin gene, and five of the $\beta$ - globin gene. Different combinations of these genes are expressed during development of the human, resulting in the hemoglobin having different binding properties for oxygen. The presence of a copy of a gene can therefore have benefits. Biologically the results of gene duplications can be to supply genetic material capable of:
(i) Acquiring lethal mutation(s). If two copies of a gene are present in a genome, and one happens to be mutated such that it's protein product is no longer functional having the second copy means that this

probability of harmful mutation is reduced.
(ii) Evolving new functions. If a mutation is favourable it may result in an altered protein product, perhaps with new functionality, without changing the function of the original gene's protein product.
(iii) Duplicating a protein in the cell, thus producing more of it. The increased presence of a protein in a cell may have some biological effects.

Goldberg [Goldberg 89] stated that using operators such as duplication in a variable-length genotype enables them to solve problems by combining relatively short, well-tested building blocks to form longer, more complex strings that increasingly cover all features of a problem.

The duplication operator used in GE is a multiple gene duplication operator. The number of genes to be duplicated is picked randomly. The duplicated genes are placed into the position of the last gene on the chromosome, if the (X * X) part of the expression were duplicated, this could result in a useful extension to the genome, especially if a multiplication operator happend to be between the original expressions genes and the duplicated one. Gene duplication in GE is essentially analogous to producing more copies of a gene(s) to increase the presence of a protein(s) in the cell.

next up previous
Next: Pruning Up: Operators Previous: Operators
Red Hat Linux User