FOLD PREDICTION BY A HIERARCHY OF SEQUENCE, THREADING, AND MODELING METHODS

Several fold recognition algorithms are compared to each other in term s of prediction accuracy and significance. It is shown that on standar d benchmarks, hybrid methods, which combine scoring based on sequence- sequence and sequence-structure matching, surpass both sequence and th reading methods in the number of accurate predictions. However, the se quence similarity contributes most to the prediction accuracy. This st rongly argues that most examples of apparently nonhomologous proteins with similar folds are actually related by evolution. While disappoint ing from the perspective of the fundamental understanding of protein f olding, this adds a new significance to fold recognition methods as a possible first step in function prediction. Despite hybrid methods bei ng more accurate at fold prediction than either the sequence or thread ing methods, each of the methods is correct in some cases where others have failed. This partly reflects a different perspective on sequence /structure relationship embedded in various methods. To combine predic tions from different methods, estimates of significance of predictions are made for all methods. With the help of such estimates, it is poss ible to develop a ''jury'' method, which has accuracy higher than any of the single methods. Finally, building full three-dimensional models for all top predictions helps to eliminate possible false positives w here alignments, which are optimal in the one-dimensional sequences, l ead to unsolvable sterical conflicts for the full three-dimensional mo dels.