More than four decade have passed since protein folding was recognised as a "problem" inasmuch as nature can find the right conformation out of an astronomical number of possibilities in a split second using only the information contained in the sequence, while we mere mortals can look at the amino acid chain and have trouble to work out which bit goes where. During that time, a lot of effort has gone into clarifying some mechanistic principles, elucidating the role of helper proteins (molecular chaperones) in the cell, and creating algorithms that can predict structures of small proteins. For large proteins, however, folding prediction is still a hard problem.
I've written a feature covering some of the most original approaches that have in recent years been applied to the folding (prediction) problem, including the distributed computing network folding@home, the crowdsourcing project Foldit, and recent work on using the information included in evolutionary relatedness of protein families. The feature appears in Current Biology today and is freely accessible to all:
Folding research recruits unconventional help
Current Biology, Volume 22, Issue 2, R35-R38, 24 January 2012
PS In other news, my previous feature, We need to talk about nitrogen, is currently in 8th position in the Current Biology download chart.