prev next front |1 |2 |3 |4 |5 |6 |7 |8 |9 |10 |11 |12 |13 |14 |15 |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |29 |30 |review
This slide illustrates several of the sequence representations to be discussed at this meeting. There are several  papers on novel methods for discovering sequence motifs or consensus sequences.  The primary goal of the new work  is to discover conserved properties of  residue rather than conserved residues themselves.
Dynamic programming methods can introduce more flexible patterns allowing arbitrary mismatches and gaps.
Weight matrices, originally introduced by Roger Staden followed by the Blocks of Steve Henikoff and the Profiles of Mike Gribsko and Templates of Tom Blundell  permit  increasingly  robust representations of motifs, although often at the sacrifice of some precision or sensitivity. There advantage is that they permit the representation and dynamic programming of entire families of sequences rather than just a single one.
Finally, there are tutorials, papers and posters on Hidden Markov models of proteins, a clear generalization of the profile concept appropriate for the larger superfamilies.
These methods themselves can be classified along several dimensions one of which is shown on the next slide.