3D Structural Homology Detection via Unassigned Residual Dipolar Couplings

• Authors: Christopher James Langmead;Bruce Randall Donald
• Venue: CSB '03 Proceedings of the IEEE Computer Society Conference on Bioinformatics
• Year: 2003

Recognition of a protein's fold provides valuable informationabout its function. While many sequence-based homologyprediction methods exist, an important challengeremains: two highly dissimilar sequences can have similarfolds - how can we detect this rapidly, in the context ofstructural genomics? High-throughput NMR experiments,coupled with novel algorithms for data analysis, can addressthis challenge. We report an automated procedure fordetecting 3D structural homologies from sparse, unassignedprotein NMR data.Our method identifies the 3D structural models in a proteinstructural database whose geometries best fit the unassignedexperimental NMR data. It does not use sequenceinformation and is thus not limited by sequence homology.The method can also be used to confirm or refutestructural predictions made by other techniques such asprotein threading or sequence homology. The algorithmruns in O(pnk3) time, where p is the number of proteinsin the database, n is the number of residues in the targetprotein, and k is the resolution of a rotation search.The method requires only uniform 15N-labelling of the proteinand processes unassigned HN-15N residual dipolarcouplings, which can be acquired in a couple of hours.Our experiments on NMR data from 5 different proteinsdemonstrate that the method identifies closely related proteinfolds, despite low-sequence homology between the targetprotein and the computed model.