A probabilistic model for secondary structure prediction from protein chemical shifts

Abstract

Protein chemical shifts encode detailed structural information that is difficult and computationally costly to describe at a fundamental level. Statistical and machine learning approaches have been used to infer correlations between chemical shifts and secondary structure from experimental chemical shifts. These methods range from simple statistics such as the chemical shift index to complex methods using neural networks. Notwithstanding their higher accuracy, more complex approaches tend to obscure the relationship between secondary structure and chemical shift and often involve many parameters that need to be trained. We present hidden Markov models (HMMs) with Gaussian emission probabilities to model the dependence between protein chemical shifts and secondary structure. The continuous emission probabilities are modeled as conditional probabilities for a given amino acid and secondary structure type. Using these distributions as outputs of first- and second-order HMMs, we achieve a prediction accuracy of 82.3%, which is competitive with existing methods for predicting secondary structure from protein chemical shifts. Incorporation of sequence-based secondary structure prediction into our HMM improves the prediction accuracy to 84.0%. Our findings suggest that an HMM with correlated Gaussian distributions conditioned on the secondary structure provides an adequate generative model of chemical shifts.

Best Practices in Utilization of 2D-NMR Spectral Data as the Input for Chemometric Analysis in Biopharmaceutical Applications

“…30 This effort has had some success, as there are a number of reports describing the prediction of secondary structure from a protein sequence and an unassigned peak table. 31 However, by and large, greater success in utilization of chemical shift data has been achieved if a PDB structure is available. 32−35 Despite the complex relationship between structure and chemical shift, this NMR parameter is very robust combined metric of structural and chemical environments and is invariant with respect to acquisition strategy or field strength.…”

“…Much ink has been spilled in attempts to interpret or predict structural details that cause a given atom to resonate at a particular frequency . This effort has had some success, as there are a number of reports describing the prediction of secondary structure from a protein sequence and an unassigned peak table . However, by and large, greater success in utilization of chemical shift data has been achieved if a PDB structure is available. − …”

Identify five kinds of simple super-secondary structures with quadratic discriminant algorithm based on the chemical shifts

“…Therefore, some researchers have utilized chemical shift for the determination of biomolecular structures (Case, 1998;Wishart and Case, 2001). Moreover, some works have studied on protein structure prediction (Cavalli et al, 2007;Lin et al, 2012;Mao et al, 2013;Mechelke and Habeck, 2013;Mielke and Krishnan, 2003;Pastore and Saudek, 1990;Wang, 2004;Zhang et al, 2003) and protein backbone and side chain torsion angle prediction (Shen and Bax, 2013) by using chemical shifts, these results showed that chemical shift is a powerful parameter for the determination of protein structure information.…”

A magnetic biocatalyst based on mussel-inspired polydopamine and its acylation of dihydromyricetin

“…This suggests an increase of β-sheet and α-helix content upon lipase immobilization of 2.74 and 0.84 percentage points, respectively (Table 1). The existence of hydrogen bonds within secondary structure elements such as β-sheet and α-helix generally helps to maintain protein structure [42,43], which may be one of the reasons why the immobilized lipase exhibited enhanced stability and improved tolerance to organic solvents and ionic liquid [44].…”