More BLAST for the Buck

Time and money—they are the essential, yet most often, the limiting factors in people’s lives. Sequence comparisons via computer searches are certainly no different. For the sequence similarity search, time is determined by how long it takes to compare your sequence of interest to the burgeoning sequence information available, whereas matches found are the currency. Of course, not just any sequence match is worthwhile—only matches that are likely to have biological significance; your money should be worth something. To obtain such matches requires a program that is both sensitive (able to pick up relevant matches of weak similarity) and selective (provide a list of matches that are all biologically relevant). The balance between these two can greatly increase the time it takes to run a program—too high a sensitivity and it will take too much time and produce too many nonbiologically relevant sequences; too low a sensitivity and the program will run much faster, but meaningful matches may be lost. The BLAST program has been the primary method for doing database searches to detect related sequences for the purpose of classifying proteins into functional families and identifying unknown functional units. Now, a paper by Altschul et al. (1997) provides modifications to the current algorithms in the BLAST program that save the user time and provide far greater return on the investment.

The new algorithms provide three important changes to the current BLAST program. The first is that the program can run approximately two to three times faster—important in a world of rapidly expanding sequence information. The second is that the program can now allow the production of gapped alignments, which often result in detection of biologically relevant matches that were originally overlooked. The third is the development of a motif or profile searching program within the BLAST system. Motif …