Forthcoming attractions

Sequence analysis The hot topic in sequence analysis is hidden Markov models (HMMs). Rather than looking at related protein sequences as clusters of pairwise alignments, HMMs abstract from a family of related sequences a probability model that can be used to predict whether a novel sequence is a member of the family. Some nifty mathematics are needed for accurate HMMs, but once created they can be extremely sensitive to cryptic family members typically missed by conventional methods. Many resources for HMM researchers are collected at the University of California Santa Cruz (http://www.cse.ucsc.edu/research/ compbio/), and a collection of readyto-use HMMs are held at the Sanger Center in the UK (http://www.sanger. ac.uk/Pfam/), which provides an interesting alternative to databases such as PROSITE (http://expasy. hcuge.ch/sprot/prosite.html) for identifying protein sequence motifs.