HEURISTIC SEARCH IN CONSTRAINED BIPARTITE MATCHING WITH APPLICATIONS TO PROTEIN NMR BACKBONE RESONANCE ASSIGNMENT

From Raw Data to Protein Backbone Chemical Shifts Using NMRFx Processing and NMRViewJ Analysis

Assignment of the chemical shifts of the backbone atoms (HN, N, CA, CB, and C) of proteins is often a prerequisite to using NMR information in the study of proteins. These chemical shifts and their perturbations are the basis for the analysis of protein dynamics, ligand binding, and backbone conformation. They are generally assigned prior to full side-chain assignments and the determination of the complete three-dimensional molecular structure. This chapter describes the use of two software packages, NMRFx Processor and NMRViewJ, in going from raw NMR data to backbone assignments. The step-by-step procedure describes processing of the data and the use of manual and automated features of the RunAbout tool in NMRViewJ to perform the assignments.

CISA: combined NMR resonance connectivity information determination and sequential assignment

A nearly complete sequential resonance assignment is a key factor leading to successful protein structure determination via NMR spectroscopy. Assuming the availability of a set of NMR spectral peak lists, most of the existing assignment algorithms first use the differences between chemical shift values for common nuclei across multiple spectra to provide the evidence that some pairs of peaks should be assigned to sequentially adjacent amino acid residues in the target protein. They then use these connectivities as constraints to produce a sequential assignment. At various levels of success, these algorithms typically generate a large number of potential connectivity constraints, and it grows exponentially as the quality of spectral data decreases. A key observation used in our sequential assignment program, CISA, is that chemical shift residual signature information can be used to improve the connectivity determination, and thus to dramatically decrease the number of predicted connectivity constraints. Fewer connectivity constraints lead to less ambiguities in the sequential assignment. Extensive simulation studies on several large test datasets demonstrated that CISA is efficient and effective, compared to three most recently proposed sequential resonance assignment programs RANDOM, PACES, and MARS.