This review provides an exposition to the important problems of (i) structure prediction in protein folding and (ii) de novo protein design. The recent advances in protein folding are reviewed based on a classification of the approaches in comparative modeling, fold recognition, and first principles methods with and without database information. The advances towards the challenging problem of loop structure prediction and the first principles method, ASTRO-FOLD, along with the developments in the area of force-fields development have been discussed. Finally, the recent progress in the area of de novo protein design is presented with focus on template flexibility, in silico sequence selection, and successful peptide and protein designs. (c) 2005 Elsevier Ltd. All rights reserved.