Automated analysis of protein NMR assignments and structures.

A powerful feature of macromolecular structure analysis by NMR spectroscopy is its potential for automation.1,2 It has been recognized for some time that many of the interactive tasks carried out by an expert in the process of spectral analysis could, in principle, be carried out more efficiently and rapidly by computational systems. Manual methods of protein data analysis often involve using laboratoryspecific, or even user-specific, rules of interpretation and validation, which are difficult (if not impossible) to document and reproduce from one laboratory to another. In a sense, until such rules of data interpretation and validation can be standardized, the NMR structure analysis process will retain subjective aspects that limit its reproducibility and compromise its scientific value. For these reasons, a critical next step in evolving a level of scientific maturity in the field of biomolecular NMR is to establish conventions and standards of data interpretation and validation and to instantiate these standards as part of a largely automated process of data analysis. In this review we summarize recent advances in automating the processes of determining 3D structures of proteins from NMR data, with emphasis on those methods of computational and experimental NMR which have been incorporated into automated analysis platforms.