DNA-DNA hybridisation

DNA-DNA hybridization is a method in genetics to measure the degree of genetic similarity between DNA sequences. The technique is usually used to determine the genetic "distance" between two species. When several species are compared that way, the similarity values allow the species to be arranged in a phylogenetic tree.

Method

The DNA from the two species to be compared is extracted, purified and cut into short pieces (e.g., 600-800 base pairs). The DNA double strand is then separated by heating into two single strands. The single-stranded DNA is now allowed to anneal with the DNA pieces of the other species. The more similar the DNA, the more of the pieces will anneal and form a hybrid (thus the name) double strand. Strands with a high degree of similarity will bind more firmly, and require more energy to separate them: i.e. they separate when heated at a higher temperature than dissimilar strands. To assess this "melting temperature" the mixture is heated in small steps. At each step, samples are tested as to the amount of single- and double-stranded DNA. These results in a profile from which the amount of similar DNA, and thus the degree of genetic similarity, can be determined.

Pros and Cons

This technique was considered a good one since it took all possible ways of aligning the sequences into account and the melting temperature would be a good average. However it is not considered the best approach these days since sequences can be computationally aligned. There is hardly any other approach used currently, however the sequence/s used for the comparisons are the major source of contention. Not all sequences evolve at the same rate. Some are too critical and changes can be lost if the result causes loss of function of the gene product/action. Finding the corresponding genes in more distant organisms can also be difficult. Some approaches have considered using non-coding sequences since these are not believed to be affected by evolution and all mutations in them would be retained faithfully. If one assumes a constant rate of background mutation, these mutations would indicate the age of the sequence lineage. However, these too are difficult for cross genera comparisons.

A really excellent text on these topics is Graur and Li (add reference)