Protein secondary structure prediction for a single-sequence using hidden semi-Markov models

By Z. Aydın, Yucel Altunbasak, Mark Borodovsky. 2006

118 citations recorded by the source corpus.

BackgroundThe accuracy of protein secondary structure prediction has been improving steadily towards the 88% estimated theoretical limit. There are two types of prediction algorithms: Single-sequence prediction algorithms imply that information about other (homologous) proteins is not available, while algorithms of the second type imply that information about homologous proteins is available, and use it intensively. The single-sequence algorithms could make an important contribution to studies of proteins with no detected homologs, however the accuracy of protein secondary structure prediction from a single-sequence is not as high as when the additional evolutionary information is present.ResultsIn this paper, we further refine and extend the hidden semi-Markov model (HSMM) initially considered in the BSPSS algorithm. We introduce an improved residue dependency model by considering the patterns of statistically significant amino acid correlation at structural segment borders. We also derive models that specialize on different sections of the dependency structure and incorporate them into HSMM. In addition, we implement an iterative training method to refine estimates of HSMM parameters. The three-state-per-residue accuracy and other accuracy measures of the new method, IPSSP, are shown to be comparable or better than ones for BSPSS as well as for PSIPRED, tested under the single-sequence condition.ConclusionsWe have shown that new dependency models and training methods bring further improvements to single-sequence protein secondary structure prediction. The results are obtained under cross-validation conditions using a dataset with no pair of sequences having significant sequence similarity. As new sequences are added to the database it is possible to augment the dependency structure and obtain even higher accuracy. Current and future advances should contribute to the improvement of function prediction for orphan proteins inscrutable to current similarity search methods.

Subfields: Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

1. Jinbo Xu
2. Sheng Wang
3. Jianlin Cheng
4. Zhiyong Wang
5. Jianzhu Ma
6. Feng Zhao
7. Jesse Eickholt
8. Badri Adhikari

Citations in the corpus

• A Deep Learning Network Approach to ab initio Protein Secondary Structure Prediction (2014, methodology): “Neural networks have been effectively used in a variety of prediction algorithms, including speech and facial recognition as well as the aforementioned usage in protein secondary structure prediction [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [24], [33], [34], [35], [36]”
• Protein 8-class secondary structure prediction using conditional neural fields (2010, methodology): “Hidden Markov model (HMM) [18] is capable of describing this relationship [19, 20], but it is challenging for HMM to model the complex nonlinear relationship between input protein features and SS, especially when a large amount of heterogeneous protein features are available.”
• Learning sparse models for a dynamic Bayesian network classifier of protein secondary structure (2011, methodology): “Motivated by previous work [5,56], we make our predictions by combining the results from multiple DBN models.”
• Prediction of protein structural classes for low-homology sequences based on predicted secondary structure (2010, unknown): “Any improved secondary structure prediction [33-35] would generally lead our method to higher prediction accuracy.”
• Knowledge-based machine learning methods for macromolecular 3D structure prediction (2016, unknown): “Hidden Markov model (HMM) (Rabiner, 1989) is capable of describing this relationship (Asai et al, 1993; Aydin et al 2006), but it is challenging for HMM to model the complex nonlinear relationship between input protein features and SS, especially when a large amount of heterogeneous protein…”
• Hidden semi-Markov models (2010, unknown): “The HSMM has been applied to thirty areas, such as human activity recognition [38,55,56,79,80,119,120,128–131,140, 197,200,205,206], handwriting recognition or printed text recognition [4,15,16,27,32–35,95,96,163], network traffic characterization and anomaly detection [105,111,116,117,139,153,172,1”
• Improving protein secondary structure prediction based on short subsequences with local structure similarity (2010, background): “The estimated theoretical limit of the accuracy of secondary structure assignment from the experimentally determined 3D structure is 88% of the Q3 accuracy [5,24], which is deemed the upper bound for secondary structure prediction.”