Tuesday, October 1, 2002

Proteomics and bioinfomatics took a big step forward today. One very useful ability would be to predict protein structure, just from primary sequence information. This should be possible since the primary sequence is the information a protein uses to fold itself into its 3D stucture. The problem is the rules are complex. There are 20 different amino acids and the average protein is 300 amino acids long. We need a simple starting point -- a very small protein. In comes exendin-4, a Gila Monster saliva protein (weird). It is 39 amino acids long and they were able to build 20 amino acid long synthetic proteins based on it. The key is that this small protein still has structure. Most proteins this small are random coils in solution. Neidigh et. al. were able to correctly predict its structure from just the sequence, and they verified their prediction to be 97% correct. 9:09:01 AM    comment []