The Internet-accessible Profile Server at the SDSC has been an invaluable resource for my group's biocomputational research. Our ability to detect distant similarities between protein sequences is heightened considerably by the use of sensitive sequence profiles pioneered by Dr. Gribskov. The rationale is simple: the explosive growth of sequence databases-driven by varied genome projects-poses a great challenge to the biologist's repertoire of computational tools. It has become routine to plumb these databases for sequence homologs of cloned genes as a means to quickly assign structural or functional features to the new protein chain. However, most search algorithms are limited in their ability to efficiently cluster sequences that bear faint, fragmentary similarities to the probe chain; these relationships often provide the most valuable insights into the evolutionary design of a novel protein. We rely on the iterative buildup of sequence alignments -often aided by pattern-matching routines such as the NBCR's CAP and MOST programs-and the derivation of sequence profiles (and HMMs -Hidden Markov Models) to significantly extend the reach of database searches. My group's research delves into the structure and evolution of protein families that may be related by peculiar sequence motifs or three-dimensional folds, and participate in signaling pathways downstream of transmembrane receptors. Experimentally, we often seek homologs of our vertebrate signaling pathways in model organisms-nematodes, drosophila, yeast or even bacteria-and this often necessitates the identification of proteins that have significantly diverged in sequence-profiles are a powerful tool that ease this task. The SDSC Profile Server is uniquely configured to increase the sensitivity of profile searches by offering a 'weighting' feature that corrects for a skewed distribution of sequences in a multiple alignment file, and offers users the choice between two profile types ('Average' and 'Evolutionary'); the Server also provides a scan of the constructed profile against the latest Swissprot database release. The Server is very reliable, fast and widely accessible to researchers that have an Internet connection. Because of its ease of use and powerful applicability, I have found it an essential site to recommend to scientists, including new postdoctoral fellows or graduate students in my group or in neighboring labs at the Institute, as a tool that they should use.
Showing the most recent 10 out of 270 publications
