Sequence similarity searches on the World Wide Web.

If you have just determined the sequence of an interesting bit of DNA, one of the first questions you are likely to ask yourself is “has anybody else seen anything like this before?” Fortunately, there has been a very successful international effort to collect all known sequences (both DNA and protein) into databases so that they can be searched. These databases (and searching tools) are freely available on the Web. The problem is that these databases are HUGE and, as a result, you must compare your sequence with the vast number of other sequences. Sequence comparison is the most powerful and most reliable method of answering biological questions about the evolutionary relationships between genes. Recent improvements in the statistical methods used by similarity searching computer programs allow the biologist to make conclusions with a high level of certainty. Now that the entire genomes of some organisms have been completely sequenced, it is possible to use similarity-search methods to provide definitive answers to questions such as “Is there a copy of this interesting gene in this organism?” A database search is a similarity search, but it is frequently— although incorrectly—referred to as homology searching. The term “homology” implies a common evolutionary relationship between two traits. Just because two sequences share a stretch of nearly identical nucleotides (or amino acids [aa]) does not mean that they are directly descended from a common ancestor. Homologous proteins share a common 3-dimensional structure, while proteins with a chance similarity of aa sequence do not. Over evolutionary time, two homologous proteins may diverge to the point that sequence similarity cannot be detected, yet they may still retain their common structure. Homologous proteins generally have similar biological functions, but this is not a rigorous requirement; conversely, discovery of homology is not proof of function. Of course, a high level of similarity is a strong indication of homology. As a rule of thumb, 25% identity over a stretch of 100 aa can be considered to be good evidence of common ancestry for two sequences. Homologous sequences are usually similar over an entire sequence, or sometimes just over one functional domain, but it is never correct to state that two sequences are “50% homologous.” Homology is an all or nothing decision. Be careful in asserting homology between short regions of two sequences; matches that are more than 50% identical in a 20–40 aa region can occur by chance. At present, the two most popular programs for similarity searching are Basic Local Alignment Search Tool (BLAST) (1) and FASTA (3). They use similar approaches to reach similar answers; although, there are some subtle differences. The two programs will generally give slightly different results: in terms of which sequences are found to be similar, their relative rankings by similarity score and the statistical significance that is assigned to those similarity scores. It is generally thought that FASTA is more sensitive for DNA-DNA comparisons, but BLAST is usually faster. The choice of which program to use is often governed by factors of convenience rather than mathematical rigor; but to be thorough, it is best to perform searches with both programs. For most researchers, the Web has become the best method to make similarity searches. Web servers are available to everyone, they have the most current data, they are usually fast and your computer is not tied up while your search is computing. However, the addresses and the services offered by various Web servers (See BioBit) are subject to change without warning, so it is wise to have some alternatives lined up when your computing is dependent on the kindness of strangers. Protein similarity searches can find much more distant similarities than comparisons of DNA sequences (ca. 2.5 billion vs. 100 million years of evolutionary divergence). This is true for several