Secondary structure prediction with support vector machines

Motivation: A new method that uses support vector machines (SVMs) to predict protein secondary structure is described and evaluated. The study is designed to develop a reliable prediction method using an alternative technique and to investigate the applicability of SVMs to this type of bioinformatics problem. Methods: Binary SVMs are trained to discriminate between two structural classes. The binary classifiers are combined in several ways to predict multi-class secondary structure. Results: The average three-state prediction accuracy per protein (Q3) is estimated by cross-validation to be 77.07 ± 0.26% with a segment overlap (Sov) score of 73.32 ± 0.39%. The SVM performs similarly to the 'state-of-the-art' PSIPRED prediction method on a non-homologous test set of 121 proteins despite being trained on substantially fewer examples. A simple consensus of the SVM, PSIPRED and PROFsec achieves significantly higher prediction accuracy than the individual methods. Availability: The SVM classifier is available from the authors. Work is in progress to make the method available on-line and to integrate the SVM predictions into the PSIPRED server.

Benchmarking secondary structure prediction for fold recognition

If secondary structure predictions are to be incorporated into fold recognition methods, an assessment of the effect of specific types of errors in predicted secondary structures on the sensitivity of fold recognition should be carried out. Here, we present a systematic comparison of different secondary structure prediction methods by measuring frequencies of specific types of error. We carry out an evaluation of the effect of specific types of error on secondary structure element alignment (SSEA), a baseline fold recognition method. The results of this evaluation indicate that missing out whole helix or strand elements, or predicting the wrong type of element, is more detrimental than predicting the wrong lengths of elements or overpredicting helix or strand. We also suggest that SSEA scoring is an effective method for assessing accuracy of secondary structure prediction and perhaps may also provide a more appropriate assessment of the “usefulness” and quality of predicted secondary structure, if secondary structure alignments are to be used in fold recognition.

Protein structure prediction servers at University College London

A number of state-of-the-art protein structure prediction servers have been developed by researchers working in the Bioinformatics Unit at University College London. The popular PSIPRED server allows users to perform secondary structure prediction, transmembrane topology prediction and protein fold recognition. More recent servers include DISOPRED for the prediction of protein dynamic disorder and DomPred for domain boundary prediction.

The PSIPRED protein structure prediction server

The PSIPRED protein structure prediction server allows users to submit a protein sequence, perform a prediction of their choice and receive the results of the prediction both textually via e-mail and graphically via the web. The user may select one of three prediction methods to apply to their sequence: PSIPRED, a highly accurate secondary structure prediction method; MEMSAT 2, a new version of a widely used transmembrane topology prediction method; or GenTHREADER, a sequence profile based fold recognition method.

Statistical prediction of single-stranded regions in RNA secondary structure and application to predicting effective antisense target sites and beyond

Single-stranded regions in RNA secondary structure are important for RNA–RNA and RNA–protein interactions. We present a probability profile approach for the prediction of these regions based on a statistical algorithm for sampling RNA secondary structures. For the prediction of phylogenetically-determined single-stranded regions in secondary structures of representative RNA sequences, the probability profile offers substantial improvement over the minimum free energy structure. In designing antisense oligonucleotides, a practical problem is how to select a secondary structure for the target mRNA from the optimal structure(s) and many suboptimal structures with similar free energies. By summarizing the information from a statistical sample of probable secondary structures in a single plot, the probability profile not only presents a solution to this dilemma, but also reveals ‘well-determined’ single-stranded regions through the assignment of probabilities as measures of confidence in predictions. In antisense application to the rabbit β-globin mRNA, a significant correlation between hybridization potential predicted by the probability profile and the degree of inhibition of in vitro translation suggests that the probability profile approach is valuable for the identification of effective antisense target sites. Coupling computational design with DNA–RNA array technique provides a rational, efficient framework for antisense oligonucleotide screening. This framework has the potential for high-throughput applications to functional genomics and drug target validation.

Prediction of protein continuum secondary structure with probabilistic models based on NMR solved structures

Background: The structure of proteins may change as a result of the inherent flexibility of some protein regions. We develop and explore probabilistic machine learning methods for predicting a continuum secondary structure, i.e. assigning probabilities to the conformational states of a residue. We train our methods using data derived from high-quality NMR models. Results: Several probabilistic models not only successfully estimate the continuum secondary structure, but also provide a categorical output on par with models directly trained on categorical data. Importantly, models trained on the continuum secondary structure are also better than their categorical counterparts at identifying the conformational state for structurally ambivalent residues. Conclusion: Cascaded probabilistic neural networks trained on the continuum secondary structure exhibit better accuracy in structurally ambivalent regions of proteins, while sustaining an overall classification accuracy on par with standard, categorical prediction methods.

Rapid protein domain assignment from amino acid sequence using predicted secondary structure

The elucidation of the domain content of a given protein sequence in the absence of determined structure or significant sequence homology to known domains is an important problem in structural biology. Here we address how successfully the delineation of continuous domains can be accomplished in the absence of sequence homology using simple baseline methods, an existing prediction algorithm (Domain Guess by Size), and a newly developed method (DomSSEA). The study was undertaken with a view to measuring the usefulness of these prediction methods in terms of their application to fully automatic domain assignment. Thus, the sensitivity of each domain assignment method was measured by calculating the number of correctly assigned top scoring predictions. We have implemented a new continuous domain identification method using the alignment of predicted secondary structures of target sequences against observed secondary structures of chains with known domain boundaries as assigned by Class Architecture Topology Homology (CATH). Taking top predictions only, the success rate of the method in correctly assigning domain number to the representative chain set is 73.3%. The top prediction for domain number and location of domain boundaries was correct for 24% of the multidomain set (±20 residues). These results have been put into context in relation to the results obtained from the other prediction methods assessed

Electrochemical genosensors for the detection of Bonamia parasite. Selection of single strand-DNA (ssDNA) probes by simulation of the secondary structure folding

A post-PCR nucleic acid work by comparing experimental data, from electrochemical genosensors, and bioinformatics data, derived from the simulation of the secondary structure folding and prediction of hybridisation reaction, was carried out in order to rationalize the selection of ssDNA probes for the detection of two Bonamia species, B. exitiosa and B. ostreae, parasites of Ostrea edulis.Six ssDNA probes (from 11 to 25 bases in length, 2 thiolated and 4 biotinylated) were selected within different regions of B. ostreae and B. exitiosa PCR amplicons (300 and 304 bases, respectively) with the aim to discriminate between these parasite species. ssDNA amplicons and probes were analyzed separately using the "Mfold Web Server" simulating the secondary structure folding behaviour. The hybridisation of amplicon-probe was predicted by means of "Dinamelt Web Server". The results were evaluated considering the number of hydrogen bonds broken and formed in the simulated folding and hybridisation process, variance in gaps for each sequence and number of available bases. In the experimental part, thermally denatured PCR products were captured at the sensor interface via sandwich hybridisation with surface-tethered probes (thiolated probes) and biotinylated signalling probes. A convergence between analytical signals and simulated results was observed, indicating the possibility to use bioinformatic data for ssDNA probes selection to be incorporated in genosensors. (C) 2011 Elsevier B.V. All rights reserved.

Evolution of the secondary structure of the rRNA internal transcribed spacer 2 (ITS2) in hard ticks (Ixodidae, Arthropoda)

ITS2 sequences are used extensively in molecular taxonomy and population genetics of arthropods and other animals yet little is known about the molecular evolution of ITS2. We studied the secondary structure of ITS2 in species from each of the six main lineages of hard ticks (family Ixodidae). The ITS2 of these ticks varied in length from 679 bp in Ixodes scapularis to 1547 bp in Aponomma concolor. Nucleotide content varied also: the ITS2 of ticks from the Prostriata lineage (Ixodes spp.) had 46-49% GC whereas ITS2 sequences of ticks from the Metastriata lineage (all other hard ticks) had 61-62% GC. Despite variation in nucleotide sequence, the secondary structure of the ITS2 of all of these ticks apparently has five domains. Stems 1, 3, 4 and 5 of this secondary structure were obvious in all of the species studied. However, stem 2 was not always obvious despite the fact that it is flanked by highly conserved sequence motifs in the adjacent stems, stems 1 and 3. The ITS2 of hard ticks has apparently evolved mostly by increases and decreases in length of the nucleotide sequences, which caused increases, and decreases in the length of stems of the secondary structure. This is most obvious when stems of the secondary structures of the Prostriata (Ixodes spp.) are compared to those of the Metastriata (all other hard ticks). Increases in the size of the ITS2 may have been caused by replication slippage which generated large repeats, like those seen in Haemaphysalis humerosa and species from the Rhipicepalinae lineage, and the small repeats found in species from the other lineages of ticks.

Secondary structure of the third extracellular loop responsible for ligand selectivity of a mammalian gonadotropin-releasing hormone receptor

The extracellular loop 3 (ECL3) of the mammalian gonadotropin-releasing hormone receptor (GnRH-R) contains an acidic amino acid (Glu(301) in the mouse GnRH-R,) that confers agonist selectivity for Are in mammalian GnRH. It is proposed that a specific conformation of ECL3 is necessary to orientate the carboxyl side chain of the acidic residue for interaction with Arg(8) of GnRH, which is supported by decreased affinity for Arg(8) GnRH but not Gln(8) GnRH when an adjacent Pro is mutated to Ala. To probe the structural contribution of the loop domain to the proposed presentation of the carboxyl side chain, we synthesized a model peptide (CGPEMLNRVSEPGC) representing residues 293-302 of mouse ECL3, where Cys and Gly residues are added symmetrically at the N and C termini, respectively, allowing the introduction of a disulfide bridge to simulate the distances at which the ECL3 is tethered to the transmembrane domains 6 and 7 of the receptor. The ability of the ECL3 peptide to bind GnRH with low affinity was demonstrated by its inhibition of GnRH stimulation of inositol phosphate production in cells expressing the GnRH-R. The CD bands of the ECL3 peptides exhibited a superposition of predominantly unordered structure and partial contributions from beta-sheet structure. Likewise, the analysis of the amide I and amide III bands from micro-Raman and FT Raman experiments revealed mainly unordered conformations of the cyclic and of the linear peptide. NMR data demonstrated the presence of a beta-hairpin among an ensemble of largely disordered structures in the cyclic peptide. The location of the turn linking the two strands of the hairpin was assigned to the three central residues L-296, N-297, and R-298. A small population of structured species among an ensemble of predominantly random coil conformation suggests that the unliganded receptor represents a variety of structural conformers, some of which have the potential to make contacts with the ligand. We propose a mechanism of receptor activation whereby binding of the agonist to the inactive receptor state induces and stabilizes a particular structural state of the loop domain, leading to further conformational rearrangements across the transmembrane domain and signal propagating interaction with G proteins. Interaction of the Glu(301) of the receptor with Arg(8) of GnRH induces a folded configuration of the ligand. Our proposal thus suggests that conformational changes of both ligand and receptor result from this interaction.

Elastic properties and secondary structure formation of single-stranded DNA at monovalent and divalent salt conditions

Single-stranded DNA (ssDNA) plays a major role in several biological processes. It is therefore of fundamental interest to understand how the elastic response and the formation of secondary structures are modulated by the interplay between base pairing and electrostatic interactions. Here we measure force-extension curves (FECs) of ssDNA molecules in optical tweezers set up over two orders of magnitude of monovalent and divalent salt conditions, and obtain its elastic parameters by fitting the FECs to semiflexible models of polymers. For both monovalent and divalent salts, we find that the electrostatic contribution to the persistence length is proportional to the Debye screening length, varying as the inverse of the square root of cation concentration. The intrinsic persistence length is equal to 0.7 nm for both types of salts, and the effectivity of divalent cations in screening electrostatic interactions appears to be 100-fold as compared with monovalent salt, in line with what has been recently reported for single-stranded RNA. Finally, we propose an analysis of the FECs using a model that accounts for the effective thickness of the filament at low salt condition and a simple phenomenological description that quantifies the formation of non-specific secondary structure at low forces.

Elastic properties and secondary structure formation of single-stranded DNA at monovalent and divalent salt conditions

Single-stranded DNA (ssDNA) plays a major role in several biological processes. It is therefore of fundamental interest to understand how the elastic response and the formation of secondary structures are modulated by the interplay between base pairing and electrostatic interactions. Here we measure force-extension curves (FECs) of ssDNA molecules in optical tweezers set up over two orders of magnitude of monovalent and divalent salt conditions, and obtain its elastic parameters by fitting the FECs to semiflexible models of polymers. For both monovalent and divalent salts, we find that the electrostatic contribution to the persistence length is proportional to the Debye screening length, varying as the inverse of the square root of cation concentration. The intrinsic persistence length is equal to 0.7 nm for both types of salts, and the effectivity of divalent cations in screening electrostatic interactions appears to be 100-fold as compared with monovalent salt, in line with what has been recently reported for single-stranded RNA. Finally, we propose an analysis of the FECs using a model that accounts for the effective thickness of the filament at low salt condition and a simple phenomenological description that quantifies the formation of non-specific secondary structure at low forces.

Reduced ceria nanofilms from structure prediction

Experimentally, Ce2O3 films are used to study cerium oxide in its fully or partially reduced state, as present in many applications. We have explored the space of low energy Ce2O3 nanofilms using structure prediction and density functional calculations, yielding more than 30 distinct nanofilm structures. First, our results help to rationalize the roles of thermodynamics and kinetics in the preparation of reduced ceria nanofilms with different bulk crystalline structures (e.g. A-type or bixbyite) depending on the support used. Second, we predict a novel, as yet experimentally unresolved, nanofilm which has a structure that does not correspond to any previously reported bulk A2B3 phase and which has an energetic stability between that of A-type and bixbyite. To assist identification and fabrication of this new Ce2O3 nanofilm we calculate some observable properties and propose supports for its epitaxial growth.