A Deep Learning Approach for Prediction of Protein Secondary Structure

The secondary structure of a protein is critical for establishing a link between the protein primary and tertiary structures.For this reason,it is important to design methods for accurate protein secondary structure prediction.Most of the existing computational techniques for protein structural and functional prediction are based onmachine learning with shallowframeworks.Different deep learning architectures have already been applied to tackle protein secondary structure prediction problem.In this study,deep learning based models,i.e.,convolutional neural network and long short-term memory for protein secondary structure prediction were proposed.The input to proposed models is amino acid sequences which were derived from CulledPDB dataset.Hyperparameter tuning with cross validation was employed to attain best parameters for the proposed models.The proposed models enables effective processing of amino acids and attain approximately 87.05%and 87.47%Q3 accuracy of protein secondary structure prediction for convolutional neural network and long short-term memory models,respectively.

FTIR Analysis of Protein Secondary Structure in Cheddar Cheese during Ripening

Proteolysis is one of the most important biochemical reactions during cheese ripening.Studies on the secondary structure of proteins during ripening would be helpful for characterizing protein changes for assessing cheese quality.Fourier transform infrared spectroscopy(FTIR),with self-deconvolution,second derivative analysis and band curve-fitting,was used to characterize the secondary structure of proteins in Cheddar cheese during ripening.The spectra of the amide I region showed great similarity,while the relative contents of the secondary structures underwent a series of changes.As ripening progressed,the α-helix content decreased and the β-sheet content increased.This structural shift was attributed to the strengthening of hydrogen bonds that resulted from hydrolysis of caseins.In summary,FTIR could provide the basis for rapid characterization of cheese that is undergoing ripening.

HMM in Predicting Protein Secondary Structure

We introduced a new method---duration Hidden Markov Model (dHMM) to predicate the secondary structure of Protein. In our study, we divide the basic second structure of protein into three parts: H (a-Helix), E (B-sheet) and O (others, include coil and turn). HMM is a kind of probabilistic model which more thinking of the interaction between adjacent amino acids (these interaction were represented by transmit probability), and we use genetic algorithm to determine the model parameters. After improving on the model and fixed on the parameters of the model, we write a program HMMPS. Our example shows that HMM is a nice method for protein secondary structure prediction.

Protein Secondary Structure Prediction with Dynamic Self-Adaptation Combination Strategy Based on Entropy

The algorithm based on combination learning usually is superior to a singleclassification algorithm on the task of protein secondary structure prediction. However,the assignment of the weight of the base classifier usually lacks decision-makingevidence. In this paper, we propose a protein secondary structure prediction method withdynamic self-adaptation combination strategy based on entropy, where the weights areassigned according to the entropy of posterior probabilities outputted by base classifiers.The higher entropy value means a lower weight for the base classifier. The final structureprediction is decided by the weighted combination of posterior probabilities. Extensiveexperiments on CB513 dataset demonstrates that the proposed method outperforms theexisting methods, which can effectively improve the prediction performance.

A Hybrid Ant Colony Optimization for the Prediction of Protein Secondary Structure

Based on the concept of ant colony optimization and the idea of population in genetic algorithm, a novel global optimization algorithm, called the hybrid ant colony optimization （HACO）, is proposed in this paper to tackle continuous-space optimization problems. It was compared with other well-known stochastic methods in the optimization of the benchmark functions and was also used to solve the problem of selecting appropriate dilation efficiently by optimizing the wavelet power spectrum of the hydrophobic sequence of protein, which is the key step on using continuous wavelet transform （CWT） to predict a-helices and connecting peptides.

Ensemble Machine Learning to Enhance Q8 Protein Secondary Structure Prediction

Protein structure prediction is one of the most essential objectives practiced by theoretical chemistry and bioinformatics as it is of a vital importance in medicine,biotechnology and more.Protein secondary structure prediction(PSSP)has a significant role in the prediction of protein tertiary structure,as it bridges the gap between the protein primary sequences and tertiary structure prediction.Protein secondary structures are classified into two categories:3-state category and 8-state category.Predicting the 3 states and the 8 states of secondary structures from protein sequences are called the Q3 prediction and the Q8 prediction problems,respectively.The 8 classes of secondary structures reveal more precise structural information for a variety of applications than the 3 classes of secondary structures,however,Q8 prediction has been found to be very challenging,that is why all previous work done in PSSP have focused on Q3 prediction.In this paper,we develop an ensemble Machine Learning(ML)approach for Q8 PSSP to explore the performance of ensemble learning algorithms compared to that of individual ML algorithms in Q8 PSSP.The ensemble members considered for constructing the ensemble models are well known classifiers,namely SVM(Support Vector Machines),KNN(K-Nearest Neighbor),DT(Decision Tree),RF(Random Forest),and NB(Naïve Bayes),with two feature extraction techniques,namely LDA(Linear Discriminate Analysis)and PCA(Principal Component Analysis).Experiments have been conducted for evaluating the performance of single models and ensemble models,with PCA and LDA,in Q8 PSSP.The novelty of this paper lies in the introduction of ensemble learning in Q8 PSSP problem.The experimental results confirmed that ensemble ML models are more accurate than individual ML models.They also indicated that features extracted by LDA are more effective than those extracted by PCA.

Sensitive and Label-Free Detection of Protein Secondary Structure by Amide Ⅲ Spectral Signals using Surface-Enhanced Raman Spectroscopy

Proteins and peptides perform a vital role in living systems, however it remains a challenge for accurate description of proteins at the molecular level. Despite that surface-enhanced Raman spectroscopy (SERS) can provide the intrinsic fingerprint information of samples with ultrahigh sensitivity, it suffers from the poor reproducibility and reliability. Herein, we demonstrate that the silver nanorod array fabricated by an oblique angle deposition method is a powerful substrate for SERS to probe the protein secondary structures without exogenous labels. With this method, the SERS signals of two typical proteins (lysozyme and cytochrome c) are successfully obtained. Additionally, by analyzing the spectral signals of the amide Ⅲ of protein backbone, the influence of concentration on the folding status of proteins has been elucidated. With the concentration increasing, the components of α-helix and β-sheet structures of lysozyme increase while the secondary structures of cytochrome c almost keep constant. The SERS method in this work offers an effective optical marker to characterize the structures of proteins.

Inspections of Mobile Phone Microwaves Effects on Proteins Secondary Structure by Means of Fourier Transform Infrared Spectroscopy

In this study the effects of microwaves on the secondary structure of three typical proteins have been investigated. A set of samples of lysozyme, bovine serum albumin and myoglobin in D2O solutions were exposed for 8 hours to mobile phone microwaves at 900 MHz at a magnetic field intensity around 16 mA/m. The relative effects on the secondary structure of the proteins were studied by means of Fourier Transform Infrared Spectroscopy. An increase of the amide I band intensity in the secondary structure of the proteins was observed after the microwaves exposure. Furthermore, a weak shift of the amide I mode of bovine serum albumin and a heavier shift of the amide I of myoglobin occurred after the exposure. In addition, a clear increasing of the β-sheet components with respect to the α-helix content was observed in the spectra of bovine serum albumin and myoglobin after the exposure, suggesting the hypothesis of the formation of aggregates.

Effects of water deficit and high N fertilization on wheat storage protein synthesis,gluten secondary structure,and breadmaking quality

The content and composition of wheat storage proteins are the major determinants of dough rheological properties and breadmaking quality and are influenced by cultivation conditions.This study aimed to investigate the effects of water deficit and high N-fertilizer application on wheat storage protein synthesis,gluten secondary structure,and breadmaking quality.Reverse-phase ultrahigh-performance liquid chromatography analysis showed that storage protein and gluten macropolymer accumulation was promoted under both independent applications and a combination of water-deficit and high N-fertilizer treatments.Fourier-transform infrared spectroscopy showed that water deficit and high N-fertilizer treatments generally improved protein secondary structure formation and lipid accumulation,and reduced flour moisture.In particular,high N-fertilizer application increasedβ-sheet content by 10.4%and the combination of water-deficit and high N-fertilizer treatments increased random coil content by 7.6%.These changes in gluten content and secondary structure led to improved dough rheological properties and breadmaking quality,including superior loaf internal structure,volume,and score.Our results demonstrate that moderately high N-fertilizer application under drought conditions can improve gluten accumulation,gluten secondary structure formation,and baking quality.

Study on the secondary structure and hydration effect of human serum albumin under acidic pH and ethanol perturbation with IR/NIR spectroscopy

Human serum albumin(HSA)is the most abundant protein in plasma and plays an essential physiological role in the human body.Ethanol precipitation is the most widely used way to obtain HSA,and pH and ethanol are crucial factors affecting the process.In this study,infrared(IR)spectroscopy and near-infrared(NIR)spectroscopy in combination with chemometrics were used to investigate the changes in the secondary structure and hydration of HSA at acidic pH(5.6-3.2)and isoelectric pH when ethanol concentration was varied from 0%to 40%as a perturbation.IR spectroscopy combined with the two-dimensional correlation spectroscopy(2DCOS)analysis for acid pH system proved that the secondary structure of HSA changed significantly when pH was around 4.5.What's more,the IR spectroscopy and 2DCOS analysis showed different secondary structure forms under different ethanol concentrations at the isoelectric pH.For the hydration effect analysis,NIR spectroscopy combined with the McCabe-Fisher method and aquaphotomics showed that the free hydrogen-bonded water fluctuates dynamically,with ethanol at 0-20%enhancing the hydrogen-bonded water clusters,while weak hydrogen-bonded water clusters were formed when the ethanol concentration increased continuously from 20%to 30%.These measurements provide new insights into the structural changes and changes in the hydration behavior of HSA,revealing the dynamic process of protein purification,and providing a theoretical basis for the selection of HSA alcoholic precipitation process parameters,as well as for further studies of complex biological systems.

Impact of hot alkali modification conditions on secondary structure of peanut protein and embedding rate of curcumin

This study aimed to modify isolated and extracted peanut protein with hot alkali to study the impact of pH,heating temperature,processing time and other alkali liquor conditions on the molecular structure of the peanut.Curcumin was loaded in modified peanut protein.The results of the study are as follows:Within the alkaline range of 8<pH<12,the percentage of amino acid residue(AAR)and-turns first increased and then decreased with the increasing pH,and the percentage of AAR reached a maximum 5.21±0.33%when the pH was 11(p<0.01).The percentage of˛-helices andβ-sheets gradually decreased with increasing pH,while that of random coils gradually increased with increasing pH,reaching a maximum 11.24±0.87%when the pH was 11(p<0.05).Within the range of the heating temperature 75℃<T<95℃,the percentage of random coils andβ-sheets gradually increased with increasing heating temperature,while that of-helices and AAR gradually decreased with increasing heating temperature;they remained unchanged when the heating temperature was 90℃,and then decreased to(10.41±1.18%;p<0.01)and(4.02±2.12%;p<0.01),respectively.Within the range of 5 min<t<20 min,the percentage of random coils and AAR gradually increased with increasing heating time,while the percentage ofα-helices decreased from 11.83±1.04%to 10.75±2.34%with increased heating time(p<0.01).The optimum conditions for hot alkali modification of peanut protein as followed:heating temperature of 90℃,heating time of 20 min and a pH of alkali liquor of 11.Under these optimum conditions,the embedding rate of curcumin by the modified protein can reach 88.32±1.29%.

Neural Network Based on GA-BP Algorithm and its Application in the Protein Secondary Structure Prediction

The advantages and disadvantages of genetic algorithm and BP algorithm are introduced. A neural network based on GA-BP algorithm is proposed and applied in the prediction of protein secondary structure, which combines the advantages of BP and GA. The prediction and training on the neural network are made respectively based on 4 structure classifications of protein so as to get higher rate of predication---the highest prediction rate 75.65%,the average prediction rate 65.04%.

Using Neural Networks to Predict Secondary Structure for Protein Folding

Protein Secondary Structure Prediction (PSSP) is considered as one of the major challenging tasks in bioinformatics, so many solutions have been proposed to solve that problem via trying to achieve more accurate prediction results. The goal of this paper is to develop and implement an intelligent based system to predict secondary structure of a protein from its primary amino acid sequence by using five models of Neural Network (NN). These models are Feed Forward Neural Network (FNN), Learning Vector Quantization (LVQ), Probabilistic Neural Network (PNN), Convolutional Neural Network (CNN), and CNN Fine Tuning for PSSP. To evaluate our approaches two datasets have been used. The first one contains 114 protein samples, and the second one contains 1845 protein samples.

Identification of Secondary Structure of Extracellular Signal Regulated Kinase (ERK) Interacting Proteins and Their Domain: An in Silico Study

ERK is involved in multiple cell signaling pathways through its interacting proteins. By in silico analysis, earlier we have identified 22 putative ERK interacting proteins namely;ephrin type-B receptor 2 isoform 2 precursor (EPHB2), mitogen-activated protein kinase 1 (MAPK1), interleukin-17 receptor D precursor (IL17RD), WD repeat domain containing 83 (WDR83), tescalcin (Tesc), mitogen-activated protein kinase kinase kinase 4 (MAPP3K4), kinase suppressor of Ras2 (KSR2), mitogen-activated protein kinase kinase 6 (MAP3K6), UL16 binding protein 2 (ULBP2), UL16 binding protein 1 (ULBP1), dual specificity phosphatase 14 (DUSP14), dual specificity phosphatase 6 (DUSP6), hyaluronan-mediated motility receptor (RHAMM), kinase D interacting substrate of 220 kDa (KININS220), membrane-associated guanylate kinase (MAGI3), phosphoprotein enriched in astrocytes 15 (PEA15), typtophenyl-tRNA synthetase, cytoplasmic (WARS), dual specificity phosphatase 9 (DUSP9), mitogen-activated protein kinase kinase kinase 1 (MAP3K1), UL16 binding protein 3 (ULBP3), SLAM family member 7 isoform a precursor (SLAMMF7) and mitogen activated protein kinase kinase kinase 11 (MAP3K11) (Table 1). However, prediction of secondary structure and domain/motif present in aforementioned ERK interacting proteins is not studied. In this paper, in silico prediction of secondary structure of ERK interacting proteins was done by SOPMA and motif/domain identification using motif search. Briefly, SOPMA predicted higher random coil and alpha helix percentage in these proteins (Table 2) and motif scan predicted serine/threonine kinases active site signature and protein kinase ATP binding region in majority of ERK interacting proteins. Moreover, few have commonly dual specificity protein phosphatase family and tyrosine specific protein phosphatase domains (Table 3). Such study may be helpful to design engineered molecules for regulating ERK dependent pathways in disease condition.

Solid-State NMR Spectroscopic Approaches to Investigate Dynamics, Secondary Structure and Topology of Membrane Proteins

Solid-state NMR spectroscopy is routinely used to determine the structural and dynamic properties of both membrane proteins and peptides in phospholipid bilayers [1-26]. From the perspective of the perpetuated lipids, 2H solid-state NMR spectroscopy can be used to probe the effect of embedded proteins on the order and dynamics of the acyl chains of phospholipid bilayers [8-13]. Moreover, 31P solid-state NMR spectroscopy can be used to investigate the interaction of peptides, proteins and drugs with phospholipid head groups [11-14]. The secondary structure of 13C = O site-specific isotopically labeled peptides or proteins inserted into lipid bilayers can be probed utilizing 13C CPMAS solid-state NMR spectroscopy [15-18]. Also, solid-state NMR spectroscopic studies can be utilized to ascertain pertinent informa- tion on the backbone and side-chain dynamics of 2H- and 15N-labeled proteins, respectively, in phospholipid bilayers [19-26]. Finally, specific 15N-labeled amide sites on a protein embedded inside oriented bilayers can be used to probe the alignment of the helices with respect to the bilayer normal [2]. A brief summary of all these solid-state NMR ap- proaches are provided in this minireview.

Prediction of RNA Secondary Structure Based on Particle Swarm Optimization

A novel method for the prediction of RNA secondary structure was proposed based on the particle swarm optimization（PSO）. PSO is known to be effective in solving many different types of optimization problems and known for being able to approximate the global optimal results in the solution space. We designed an efficient objective function according to the minimum free energy, the number of selected stems and the average length of selected stems. We calculated how many legal stems there were in the sequence, and selected some of them to obtain an optimal result using PSO in the right of the objective function. A method based on the improved particle swarm optimization（IPSO） was proposed to predict RNA secondary structure, which consisted of three stages. The first stage was applied to encoding the source sequences, and to exploring all the legal stems. Then, a set of encoded stems were created in order to prepare input data for the second stage. In the second stage, IPSO was responsible for structure selection. At last, the optimal result was obtained from the secondary structures selected via IPSO. Nine sequences from the comparative RNA website were selected for the evaluation of the proposed method. Compared with other six methods, the proposed method decreased the complexity and enhanced the sensitivity and specificity on the basis of the experiment results.

Heterogeneity and Secondary Structure Analysis of 3' Untranslated Region in Classical Swine Fever Viruses

The attenuated vaccine strains of CSFV have a 12-nucleotides （nt） insertion in the 3＇-UTR of genome as compared to that of CSFV virulent strains. In this study, we found a distinct heterogeneity in the 3＇-UTR of attenuated Thiverval and HCLV strains. The longest 3＇-UTR of Thiverval strain was 259 base pairs （bp） with a 32-nt insertion, the shortest 3＇-UTR had only 233 bp with a 6-nt insertion. The longest 3＇-UTR of HCLV strain was 244 bp with a 17-nt insertion and the shortest 3＇ UTR was 235 bp with a 8-nt insertion. Compared with the published sequences of 3＇-UTR of vaccine and virulent strains, the 3＇-UTR of CSFV vaccine strains have two variable regions where insertion among the different vaccine strains were frequently found. The first is located between the second conservative TALk codon and the start of T-rich region where we found the variable length insertion in the same vaccine strain Thiveral or HCLV and the second is located between the end of T-rich region and the front of GAA eodon, however, a 4-nt deletion was found in this region in the virulent Shimen strain. These two regions may represent the ＂hot spot＂ for mutation. Modeling the secondary structures of the 3＇-UTR suggests that the T-rich insertion could result in the change of structure and free energy, thus affecting the stability of the 3＇-UTR structure. These findings will help to understand the mechanism of attenuated vaccines and improve vaccine safety, stability, and efficacy.

Electronic Structures and Alloying Behaviors of Ferrite Phases in High Co-Ni Secondary Hardened Martensitic Steels

The electronic structure of ferrite (tempered martensite phase) in high Co-Ni secondary hardened martensitic steel has been investigated. The local density of states (LOOS) of alloying elements in the steel displays the relationship between solid solubility and the shape of the LDOS. The bond order integral (BOI) between atoms in the steel shows that the directional bonding of the p orbital of Si or C leads to the brittleness of the steel. At last, ΣBOI between atoms demonstrate that C, Co, Mn, Cr, Mo, Si strengthen the alloyed steel through solid-solution effects.

Effects of β-TCP Ceramics on Intracellular Ca^(2+) Concentration,Mineralization of Osteoblast and Protein Structure

β-TCP, as one of calcium phosphates ceramics, exerts perfect biocompatibility and osteoconductivity, and is clinically used as a bone graft substitute for decades. Consequently, the effects of β-TCP ceramics on intracellular Ca2＋ concentration, mineralization of osteoblast and BSA protein structure were studied. Results showed that β-TCP could increase the intracelluar Ca2＋ concentration and mineralization of osteoblast, indicating that β-TCP ceramics could take part in the organic metabolism and the degradation product had no detrimental effect on osteoblast in vitro. Furthermore, β-TCP ceramics could increase the content of α-helix and β-pleated sheet and change BSA into more ordering structure, those changes might be favorable for the biomineralization after β-TCP ceramics implanted.

Study on Red Coat Color Gene and Prediction of the Secondary Structure in Chinese Holstein

The nucleotide sequence of the melanocortin-l-receptor （MC1R） gene was studied with the help of the polymerase chain reaction （PCR）, in which the protein structure in Chinese Holstein was predicted, and the molecular mechanism of the red coat color was investigated. Polymerase chain reaction-restriction fragment length polymorphism （PCR-RFLP） was performed to genotype the individuals. The bioinformatics and biotechnology softwares were used to predict the secondary structure of MC1R. The results showed that the EE genotype was the dominant genotype in Chinese Holstein Black and White herd, whereas, it was ee in Chinese Holstein Red and White herd. The secondary structure of the mutational MC1R protein was changed and the deletion mutation caused an earlier termination in translation, which led to the formation of the red coat color. The allele E was mainly associated with the black coat color, whereas, e was associated with red.