Prediction of transporter family from protein sequence by support vector machine approach

Machine learning and comparative genomics approaches for the discovery of xylose transporters in yeast

Background

The need to mitigate and substitute the use of fossil fuels as the main energy matrix has led to the study and development of biofuels as an alternative. Second-generation (2G) ethanol arises as one biofuel with great potential, due to not only maintaining food security, but also as a product from economically interesting crops such as energy-cane. One of the main challenges of 2G ethanol is the inefficient uptake of pentose sugars by industrial yeast Saccharomyces cerevisiae, the main organism used for ethanol production. Understanding the main drivers for xylose assimilation and identify novel and efficient transporters is a key step to make the 2G process economically viable.

Results

By implementing a strategy of searching for present motifs that may be responsible for xylose transport and past adaptations of sugar transporters in xylose fermenting species, we obtained a classifying model which was successfully used to select four different candidate transporters for evaluation in the S. cerevisiae hxt-null strain, EBY.VW4000, harbouring the xylose consumption pathway. Yeast cells expressing the transporters SpX, SpH and SpG showed a superior uptake performance in xylose compared to traditional literature control Gxf1.

Conclusions

Modelling xylose transport with the small data available for yeast and bacteria proved a challenge that was overcome through different statistical strategies. Through this strategy, we present four novel xylose transporters which expands the repertoire of candidates targeting yeast genetic engineering for industrial fermentation. The repeated use of the model for characterizing new transporters will be useful both into finding the best candidates for industrial utilization and to increase the model’s predictive capabilities.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13068-022-02153-7.

Identification of Proteins of Tobacco Mosaic Virus by Using a Method of Feature Extraction

Tobacco mosaic virus, TMV for short, is widely distributed in the global tobacco industry and has a significant impact on tobacco production. It can reduce the amount of tobacco grown by 50-70%. In this research of study, we aimed to identify tobacco mosaic virus proteins and healthy tobacco leaf proteins by using machine learning approaches. The experiment's results showed that the support vector machine algorithm achieved high accuracy in different feature extraction methods. And 188-dimensions feature extraction method improved the classification accuracy. In that the support vector machine algorithm and 188-dimensions feature extraction method were finally selected as the final experimental methods. In the 10-fold cross-validation processes, the SVM combined with 188-dimensions achieved 93.5% accuracy on the training set and 92.7% accuracy on the independent validation set. Besides, the evaluation index of the results of experiments indicate that the method developed by us is valid and robust.

TranCEP: Predicting the substrate class of transmembrane transport proteins using compositional, evolutionary, and positional information

Transporters mediate the movement of compounds across the membranes that separate the cell from its environment and across the inner membranes surrounding cellular compartments. It is estimated that one third of a proteome consists of membrane proteins, and many of these are transport proteins. Given the increase in the number of genomes being sequenced, there is a need for computational tools that predict the substrates that are transported by the transmembrane transport proteins. In this paper, we present TranCEP, a predictor of the type of substrate transported by a transmembrane transport protein. TranCEP combines the traditional use of the amino acid composition of the protein, with evolutionary information captured in a multiple sequence alignment (MSA), and restriction to important positions of the alignment that play a role in determining the specificity of the protein. Our experimental results show that TranCEP significantly outperforms the state-of-the-art predictors. The results quantify the contribution made by each type of information used.

An advanced approach to identify antimicrobial peptides and their function types for penaeus through machine learning strategies

BACKGROUND

Antimicrobial peptides (AMPs) are essential components of the innate immune system and can protect the host from various pathogenic bacteria. The marine environment is known to be one of the richest sources for AMPs. Effective usage of AMPs and their derivatives can greatly improve the immunity and breeding survival rate of aquatic products. It is highly desirable to develop computational tools for rapidly and accurately identifying AMPs and their functional types, for the purpose of helping design new and more effective antimicrobial agents.

RESULTS

In this study, we made an attempt to develop an advanced machine learning based computational approach, MAMPs-Pred, for identification of AMPs and its function types. Initially, SVM-prot 188-D features were extracted that were subsequently used as input to a two-layer multi-label classifier. In specific, the first layer is to identify whether it is an AMP by applying RF classifier, and the second layer addresses the multi-type problem by identifying the activites or function types of AMPs by applying PS-RF and LC-RF classifiers. To benchmark the methods,the MAMPs-Pred method is also compared with existing best-performing methods in literature and has shown an improved identification accuracy.

CONCLUSIONS

The results reported in this study indicate that the MAMP-Pred method achieves high performance for identifying AMPs and its functional types.The proposed approach is believed to supplement the tools and techniques that have been developed in the past for predicting AMPs and their function types.

Harnessing the evolutionary information on oxygen binding proteins through Support Vector Machines based modules

Objectives

The arrival of free oxygen on the globe, aerobic life is becoming possible. However, it has become very clear that the oxygen binding proteins are widespread in the biosphere and are found in all groups of organisms, including prokaryotes, eukaryotes as well as in fungi, plants, and animals. The exponential growth and availability of fresh annotated protein sequences in the databases motivated us to develop an improved version of “Oxypred” for identifying oxygen-binding proteins.

Results

In this study, we have proposed a method for identifying oxy-proteins with two different sequence similarity cutoffs 50 and 90%. A different amino acid composition based Support Vector Machines models was developed, including the evolutionary profiles in the form position-specific scoring matrix (PSSM). The fivefold cross-validation techniques were applied to evaluate the prediction performance. Also, we compared with existing methods, which shows nearly 97% recognition, but, our newly developed models were able to recognize almost 99.99 and 100% in both oxy-50 and 90% similarity models respectively. Our result shows that our approaches are faster and achieve a better prediction performance over the existing methods. The web-server Oxypred2 was developed for an alternative method for identifying oxy-proteins with more additional modules including PSSM, available at http://bioinfo.imtech.res.in/servers/muthu/oxypred2/home.html.

Electronic supplementary material

The online version of this article (10.1186/s13104-018-3383-9) contains supplementary material, which is available to authorized users.

Distinguishing the disease-associated SNPs based on composition frequency analysis

Single-nucleotide polymorphism (SNP) is a basical variation in genome. When SNPs occur at the binding sites of microRNA, they can influence the binding efficiency, cause a fluctuation of the mRNA in vivo, and thus arouse posttranscriptional level abnormality. Therefore, SNP has a strong correlation with diseases. Although enormous SNPs have been experimentally identified, only a tiny proportion of them are truly disease-associated SNPs (dSNPs) that relate to microRNA modification and then are involved in disease causing process. Therefore, it is important to distinguish dSNPs from the usual SNPs. Analysis here shows that composition is different between sequence segments centered by dSNP and SNP. Inspired by the composition, transition and distribution features which are meaningful and effective in characterizing proteins' sequence information, we improved and applied it to represent the frequency and physicochemical properties of a gene sequence. Binary encoding scheme was also used for further labelling four nucleic acids (A, T, C, and G). First, clustering analysis was performed to gain reasonable negative samples. Then, optimization tests were implemented on different ratios of positive vs negative samples and different feature subsets retrieved by evaluation method of F score. The optimal model constructed by random forest achieves an accuracy of more than 90% on the testing data set. Moreover, the promising results of the external validation also demonstrate the practical applicability of our method. Finally, principal component analysis on the features indicates that all features in our method gain the gross contribution to the prediction model.

Effective prediction of bacterial type IV secreted effectors by combined features of both C-termini and N-termini

Various bacterial pathogens can deliver their secreted substrates also called as effectors through type IV secretion systems (T4SSs) into host cells and cause diseases. Since T4SS secreted effectors (T4SEs) play important roles in pathogen-host interactions, identifying them is crucial to our understanding of the pathogenic mechanisms of T4SSs. A few computational methods using machine learning algorithms for T4SEs prediction have been developed by using features of C-terminal residues. However, recent studies have shown that targeting information can also be encoded in the N-terminal region of at least some T4SEs. In this study, we present an effective method for T4SEs prediction by novelly integrating both N-terminal and C-terminal sequence information. First, we collected a comprehensive dataset across multiple bacterial species of known T4SEs and non-T4SEs from literatures. Then, three types of distinctive features, namely amino acid composition, composition, transition and distribution and position-specific scoring matrices were calculated for 50 N-terminal and 100 C-terminal residues. After that, we employed information gain represent to rank the importance score of the 150 different position residues for T4SE secretion signaling. At last, 125 distinctive position residues were singled out for the prediction model to classify T4SEs and non-T4SEs. The support vector machine model yields a high receiver operating curve of 0.916 in the fivefold cross-validation and an accuracy of 85.29% for the independent test set.

Pretata: predicting TATA binding proteins with novel features and dimensionality reduction strategy

Background

It is necessary and essential to discovery protein function from the novel primary sequences. Wet lab experimental procedures are not only time-consuming, but also costly, so predicting protein structure and function reliably based only on amino acid sequence has significant value. TATA-binding protein (TBP) is a kind of DNA binding protein, which plays a key role in the transcription regulation. Our study proposed an automatic approach for identifying TATA-binding proteins efficiently, accurately, and conveniently. This method would guide for the special protein identification with computational intelligence strategies.

Results

Firstly, we proposed novel fingerprint features for TBP based on pseudo amino acid composition, physicochemical properties, and secondary structure. Secondly, hierarchical features dimensionality reduction strategies were employed to improve the performance furthermore. Currently, Pretata achieves 92.92% TATA-binding protein prediction accuracy, which is better than all other existing methods.

Conclusions

The experiments demonstrate that our method could greatly improve the prediction accuracy and speed, thus allowing large-scale NGS data prediction to be practical. A web server is developed to facilitate the other researchers, which can be accessed at http://server.malab.cn/preTata/.

Resistance gene identification from Larimichthys crocea with machine learning techniques

The research on resistance genes (R-gene) plays a vital role in bioinformatics as it has the capability of coping with adverse changes in the external environment, which can form the corresponding resistance protein by transcription and translation. It is meaningful to identify and predict R-gene of Larimichthys crocea (L.Crocea). It is friendly for breeding and the marine environment as well. Large amounts of L.Crocea's immune mechanisms have been explored by biological methods. However, much about them is still unclear. In order to break the limited understanding of the L.Crocea's immune mechanisms and to detect new R-gene and R-gene-like genes, this paper came up with a more useful combination prediction method, which is to extract and classify the feature of available genomic data by machine learning. The effectiveness of feature extraction and classification methods to identify potential novel R-gene was evaluated, and different statistical analyzes were utilized to explore the reliability of prediction method, which can help us further understand the immune mechanisms of L.Crocea against pathogens. In this paper, a webserver called LCRG-Pred is available at http://server.malab.cn/rg_lc/.

iDPF-PseRAAAC: A Web-Server for Identifying the Defensin Peptide Family and Subfamily Using Pseudo Reduced Amino Acid Alphabet Composition

Defensins as one of the most abundant classes of antimicrobial peptides are an essential part of the innate immunity that has evolved in most living organisms from lower organisms to humans. To identify specific defensins as interesting antifungal leads, in this study, we constructed a more rigorous benchmark dataset and the iDPF-PseRAAAC server was developed to predict the defensin family and subfamily. Using reduced dipeptide compositions were used, the overall accuracy of proposed method increased to 95.10% for the defensin family, and 98.39% for the vertebrate subfamily, which is higher than the accuracy from other methods. The jackknife test shows that more than 4% improvement was obtained comparing with the previous method. A free online server was further established for the convenience of most experimental scientists at http://wlxy.imu.edu.cn/college/biostation/fuwu/iDPF-PseRAAAC/index.asp. A friendly guide is provided to describe how to use the web server. We anticipate that iDPF-PseRAAAC may become a useful high-throughput tool for both basic research and drug design.

SCMMTP: identifying and characterizing membrane transport proteins using propensity scores of dipeptides

BACKGROUND

Identifying putative membrane transport proteins (MTPs) and understanding the transport mechanisms involved remain important challenges for the advancement of structural and functional genomics. However, the transporter characters are mainly acquired from MTP crystal structures which are hard to crystalize. Therefore, it is desirable to develop bioinformatics tools for the effective large-scale analysis of available sequences to identify novel transporters and characterize such transporters.

RESULTS

This work proposes a novel method (SCMMTP) based on the scoring card method (SCM) using dipeptide composition to identify and characterize MTPs from an existing dataset containing 900 MTPs and 660 non-MTPs which are separated into a training dataset consisting 1,380 proteins and an independent dataset consisting 180 proteins. The SCMMTP produced estimating propensity scores for amino acids and dipeptides as MTPs. The SCMMTP training and test accuracy levels respectively reached 83.81% and 76.11%. The test accuracy of support vector machine (SVM) using a complicated classification method with a low possibility for biological interpretation and position-specific substitution matrix (PSSM) as a protein feature is 80.56%, thus SCMMTP is comparable to SVM-PSSM. To identify MTPs, SCMMTP is applied to three datasets including: 1) human transmembrane proteins, 2) a photosynthetic protein dataset, and 3) a human protein database. MTPs showing α-helix rich structure is agreed with previous studies. The MTPs used residues with low hydration energy. It is hypothesized that, after filtering substrates, the hydrated water molecules need to be released from the pore regions.

CONCLUSIONS

SCMMTP yields estimating propensity scores for amino acids and dipeptides as MTPs, which can be used to identify novel MTPs and characterize transport mechanisms for use in further experiments.

AVAILABILITY

http://iclab.life.nctu.edu.tw/iclab_webtools/SCMMTP/.

A sequence-based two-level method for the prediction of type I secreted RTX proteins

Many Gram-negative bacteria use the type I secretion system (T1SS) to translocate a wide range of substrates (type I secreted RTX proteins, T1SRPs) from the cytoplasm across the inner and outer membrane in one step to the extracellular space. Since T1SRPs play an important role in pathogen-host interactions, identifying them is crucial for a full understanding of the pathogenic mechanism of T1SS. However, experimental identification is often time-consuming and expensive. In the post-genomic era, it becomes imperative to predict new T1SRPs using information from the amino acid sequence alone when new proteins are being identified in a high-throughput mode. In this study, we report a two-level method for the first attempt to identify T1SRPs using sequence-derived features and the random forest (RF) algorithm. At the full-length sequence level, the results show that the unique feature of T1SRPs is the presence of variable numbers of the calcium-binding RTX repeats. These RTX repeats have a strong predictive power and so T1SRPs can be well distinguished from non-T1SRPs. At another level, different from that of the secretion signal, we find that a sequence segment located at the last 20-30 C-terminal amino acids may contain important signal information for T1SRP secretion because obvious differences were shown between the corresponding positions of T1SRPs and non-T1SRPs in terms of amino acid and secondary structure compositions. Using five-fold cross-validation, overall accuracies of 97% at the full-length sequence level and 89% at the secretion signal level were achieved through feature evaluation and optimization. Benchmarking on an independent dataset, our method could correctly predict 63 and 66 of 74 T1SRPs at the full-length sequence and secretion signal levels, respectively. We believe that this study will be useful in elucidating the secretion mechanism of T1SS and facilitating hypothesis-driven experimental design and validation.

Discrimination of membrane transporter protein types using K-nearest neighbor method derived from the similarity distance of total diversity measure

Membrane transporters play crucial roles in the fundamental cellular processes of living organisms. Computational techniques are very necessary to annotate the transporter functions. In this study, a multi-class K nearest neighbor classifier based on the increment of diversity (KNN-ID) was developed to discriminate the membrane transporter types when the increment of diversity (ID) was introduced as one of the novel similarity distances. Comparisons with multiple recently published methods showed that the proposed KNN-ID method outperformed the other methods, obtaining more than 20% improvement for overall accuracy. The overall prediction accuracy reached was 83.1%, when the K was selected as 2. The prediction sensitivity achieved 76.7%, 89.1%, 80.1% for channels/pores, electrochemical potential-driven transporters, primary active transporters, respectively. Discrimination and comparison between any two different classes of transporters further demonstrated that the proposed method is a potential classifier and will play a complementary role for facilitating the functional assignment of transporters.

A consensus subunit-specific model for annotation of substrate specificity for ABC transporters

A consensus classification model was built by considering three subunit proteins individually to predict the substrate specificity of ABC transporters.

Prediction of membrane transport proteins and their substrate specificities using primary sequence information

Background

Membrane transport proteins (transporters) move hydrophilic substrates across hydrophobic membranes and play vital roles in most cellular functions. Transporters represent a diverse group of proteins that differ in topology, energy coupling mechanism, and substrate specificity as well as sequence similarity. Among the functional annotations of transporters, information about their transporting substrates is especially important. The experimental identification and characterization of transporters is currently costly and time-consuming. The development of robust bioinformatics-based methods for the prediction of membrane transport proteins and their substrate specificities is therefore an important and urgent task.

Results

Support vector machine (SVM)-based computational models, which comprehensively utilize integrative protein sequence features such as amino acid composition, dipeptide composition, physico-chemical composition, biochemical composition, and position-specific scoring matrices (PSSM), were developed to predict the substrate specificity of seven transporter classes: amino acid, anion, cation, electron, protein/mRNA, sugar, and other transporters. An additional model to differentiate transporters from non-transporters was also developed. Among the developed models, the biochemical composition and PSSM hybrid model outperformed other models and achieved an overall average prediction accuracy of 76.69% with a Mathews correlation coefficient (MCC) of 0.49 and a receiver operating characteristic area under the curve (AUC) of 0.833 on our main dataset. This model also achieved an overall average prediction accuracy of 78.88% and MCC of 0.41 on an independent dataset.

Conclusions

Our analyses suggest that evolutionary information (i.e., the PSSM) and the AAIndex are key features for the substrate specificity prediction of transport proteins. In comparison, similarity-based methods such as BLAST, PSI-BLAST, and hidden Markov models do not provide accurate predictions for the substrate specificity of membrane transport proteins. TrSSP: The Transporter Substrate Specificity Prediction Server, a web server that implements the SVM models developed in this paper, is freely available at http://bioinfo.noble.org/TrSSP.

Transporter taxonomy - a comparison of different transport protein classification schemes

Currently, there are more than 800 well characterized human membrane transport proteins (including channels and transporters) and there are estimates that about 10% (approx. 2000) of all human genes are related to transport. Membrane transport proteins are of interest as potential drug targets, for drug delivery, and as a cause of side effects and drug–drug interactions. In light of the development of Open PHACTS, which provides an open pharmacological space, we analyzed selected membrane transport protein classification schemes (Transporter Classification Database, ChEMBL, IUPHAR/BPS Guide to Pharmacology, and Gene Ontology) for their ability to serve as a basis for pharmacology driven protein classification. A comparison of these membrane transport protein classification schemes by using a set of clinically relevant transporters as use-case reveals the strengths and weaknesses of the different taxonomy approaches.

Support vector machines coupled with proteomics approaches for detecting biomarkers predicting chemotherapy resistance in small cell lung cancer

The aim of this study was to identify serum protein fingerprints of small cell lung cancer (SCLC) and potential biomarkers related to chemotherapy resistance of SCLC with surface enhanced laser desorption/ionization time of flight mass spectrometry (SELDI-TOF MS). A total of 60 SCLC patients and 48 age- and sex-matched healthy individuals were enrolled. The chemotherapy regimen was cisplatin plus etoposide. All patients received two cycles of chemotherapy. Serum protein profiles were detected using SELDI-TOF MS and the spectra were analyzed with support vector machines (SVMs). Western blotting was performed to verify the results of SELDI-TOF MS. Three top scored peaks, at m/z of 6269, 9043 and 13124 Da, were finally selected as potential biomarkers for detection of SCLC. The SVM classifier separated the SCLC from the healthy samples in the blind test, with a sensitivity of 92.4% and a specificity of 92.5%. For the 56 eligible chemotherapy patients, 4 had a complete response (7.14%), 39 patients had a partial response (69.6%), 9 patients had a stable disease (16.1%) and 4 patients had a progressive disease (7.14%). The model constructed using two protein peaks with m/z of 8830 and 10468 Da separated the chemotherapy-resistant group from the chemotherapy-sensitive group with a sensitivity of 80.0% and a specificity of 80.0%. Initial protein database searching identified 10468 Da as S100-A9 which was confirmed by western blotting. The present results suggest that the combination of SELDI-TOF MS with SVM may provide a useful means in the search for serum biomarkers for predicting chemotherapy resistance in patients with SCLC.

MISS-Prot: web server for self/non-self discrimination of protein residue networks in parasites; theory and experiments in Fasciola peptides and Anisakis allergens

Infections caused by human parasites (HPs) affect the poorest 500 million people worldwide but chemotherapy has become expensive, toxic, and/or less effective due to drug resistance. On the other hand, many 3D structures in Protein Data Bank (PDB) remain without function annotation. We need theoretical models to quickly predict biologically relevant Parasite Self Proteins (PSP), which are expressed differentially in a given parasite and are dissimilar to proteins expressed in other parasites and have a high probability to become new vaccines (unique sequence) or drug targets (unique 3D structure). We present herein a model for PSPs in eight different HPs (Ascaris, Entamoeba, Fasciola, Giardia, Leishmania, Plasmodium, Trypanosoma, and Toxoplasma) with 90% accuracy for 15 341 training and validation cases. The model combines protein residue networks, Markov Chain Models (MCM) and Artificial Neural Networks (ANN). The input parameters are the spectral moments of the Markov transition matrix for electrostatic interactions associated with the protein residue complex network calculated with the MARCH-INSIDE software. We implemented this model in a new web-server called MISS-Prot (MARCH-INSIDE Scores for Self-Proteins). MISS-Prot was programmed using PHP/HTML/Python and MARCH-INSIDE routines and is freely available at: . This server is easy to use by non-experts in Bioinformatics who can carry out automatic online upload and prediction with 3D structures deposited at PDB (mode 1). We can also study outcomes of Peptide Mass Fingerprinting (PMFs) and MS/MS for query proteins with unknown 3D structures (mode 2). We illustrated the use of MISS-Prot in experimental and/or theoretical studies of peptides from Fasciola hepatica cathepsin proteases or present on 10 Anisakis simplex allergens (Ani s 1 to Ani s 10). In doing so, we combined electrophoresis (1DE), MALDI-TOF Mass Spectroscopy, and MASCOT to seek sequences, Molecular Mechanics + Molecular Dynamics (MM/MD) to generate 3D structures and MISS-Prot to predict PSP scores. MISS-Prot also allows the prediction of PSP proteins in 16 additional species including parasite hosts, fungi pathogens, disease transmission vectors, and biotechnologically relevant organisms.

Identification of small molecule aggregators from large compound libraries by support vector machines

Small molecule aggregators non-specifically inhibit multiple unrelated proteins, rendering them therapeutically useless. They frequently appear as false hits and thus need to be eliminated in high-throughput screening campaigns. Computational methods have been explored for identifying aggregators, which have not been tested in screening large compound libraries. We used 1319 aggregators and 128,325 non-aggregators to develop a support vector machines (SVM) aggregator identification model, which was tested by four methods. The first is five fold cross-validation, which showed comparable aggregator and significantly improved non-aggregator identification rates against earlier studies. The second is the independent test of 17 aggregators discovered independently from the training aggregators, 71% of which were correctly identified. The third is retrospective screening of 13M PUBCHEM and 168K MDDR compounds, which predicted 97.9% and 98.7% of the PUBCHEM and MDDR compounds as non-aggregators. The fourth is retrospective screening of 5527 MDDR compounds similar to the known aggregators, 1.14% of which were predicted as aggregators. SVM showed slightly better overall performance against two other machine learning methods based on five fold cross-validation studies of the same settings. Molecular features of aggregation, extracted by a feature selection method, are consistent with published profiles. SVM showed substantial capability in identifying aggregators from large libraries at low false-hit rates.

Prediction of enzyme classes from 3D structure: a general model and examples of experimental-theoretic scoring of peptide mass fingerprints of Leishmania proteins

The number of protein and peptide structures included in Protein Data Bank (PDB) and Gen Bank without functional annotation has increased. Consequently, there is a high demand for theoretical models to predict these functions. Here, we trained and validated, with an external set, a Markov Chain Model (MCM) that classifies proteins by their possible mechanism of action according to Enzyme Classification (EC) number. The methodology proposed is essentially new, and enables prediction of all EC classes with a single equation without the need for an equation for each class or nonlinear models with multiple outputs. In addition, the model may be used to predict whether one peptide presents a positive or negative contribution of the activity of the same EC class. The model predicts the first EC number for 106 out of 151 (70.2%) oxidoreductases, 178/178 (100%) transferases, 223/223 (100%) hydrolases, 64/85 (75.3%) lyases, 74/74 (100%) isomerases, and 100/100 (100%) ligases, as well as 745/811 (91.9%) nonenzymes. It is important to underline that this method may help us predict new enzyme proteins or select peptide candidates that improve enzyme activity, which may be of interest for the prediction of new drugs or drug targets. To illustrate the model's application, we report the 2D-Electrophoresis (2DE) isolation from Leishmania infantum as well as MADLI TOF Mass Spectra characterization and theoretical study of the Peptide Mass Fingerprints (PMFs) of a new protein sequence. The theoretical study focused on MASCOT, BLAST alignment, and alignment-free QSAR prediction of the contribution of 29 peptides found in the PMF of the new protein to specific enzyme action. This combined strategy may be used to identify and predict peptides of prokaryote and eukaryote parasites and their hosts as well as other superior organisms, which may be of interest in drug development or target identification.

TransportTP: a two-phase classification approach for membrane transporter prediction and characterization

BACKGROUND

Membrane transporters play crucial roles in living cells. Experimental characterization of transporters is costly and time-consuming. Current computational methods for transporter characterization still require extensive curation efforts, especially for eukaryotic organisms. We developed a novel genome-scale transporter prediction and characterization system called TransportTP that combined homology-based and machine learning methods in a two-phase classification approach. First, traditional homology methods were employed to predict novel transporters based on sequence similarity to known classified proteins in the Transporter Classification Database (TCDB). Second, machine learning methods were used to integrate a variety of features to refine the initial predictions. A set of rules based on transporter features was developed by machine learning using well-curated proteomes as guides.

RESULTS

In a cross-validation using the yeast proteome for training and the proteomes of ten other organisms for testing, TransportTP achieved an equivalent recall and precision of 81.8%, based on TransportDB, a manually annotated transporter database. In an independent test using the Arabidopsis proteome for training and four recently sequenced plant proteomes for testing, it achieved a recall of 74.6% and a precision of 73.4%, according to our manual curation.

CONCLUSIONS

TransportTP is the most effective tool for eukaryotic transporter characterization up to date.

Prediction of functional class of proteins and peptides irrespective of sequence homology by support vector machines

Various computational methods have been used for the prediction of protein and peptide function based on their sequences. A particular challenge is to derive functional properties from sequences that show low or no homology to proteins of known function. Recently, a machine learning method, support vector machines (SVM), have been explored for predicting functional class of proteins and peptides from amino acid sequence derived properties independent of sequence similarity, which have shown promising potential for a wide spectrum of protein and peptide classes including some of the low- and non-homologous proteins. This method can thus be explored as a potential tool to complement alignment-based, clustering-based, and structure-based methods for predicting protein function. This article reviews the strategies, current progresses, and underlying difficulties in using SVM for predicting the functional class of proteins. The relevant software and web-servers are described. The reported prediction performances in the application of these methods are also presented.

QSAR for RNases and theoretic-experimental study of molecular diversity on peptide mass fingerprints of a new Leishmania infantum protein

The toxicity and low success of current treatments for Leishmaniosis determines the search of new peptide drugs and/or molecular targets in Leishmania pathogen species (L. infantum and L. major). For example, Ribonucleases (RNases) are enzymes relevant to several biologic processes; then, theoretical and experimental study of the molecular diversity of Peptide Mass Fingerprints (PMFs) of RNases is useful for drug design. This study introduces a methodology that combines QSAR models, 2D-Electrophoresis (2D-E), MALDI-TOF Mass Spectroscopy (MS), BLAST alignment, and Molecular Dynamics (MD) to explore PMFs of RNases. We illustrate this approach by investigating for the first time the PMFs of a new protein of L. infantum. Here we report and compare new versus old predictive models for RNases based on Topological Indices (TIs) of Markov Pseudo-Folding Lattices. These group of indices called Pseudo-folding Lattice 2D-TIs include: Spectral moments pi ( k )(x,y), Mean Electrostatic potentials xi ( k )(x,y), and Entropy measures theta ( k )(x,y). The accuracy of the models (training/cross-validation) was as follows: xi ( k )(x,y)-model (96.0%/91.7%)>pi ( k )(x,y)-model (84.7/83.3) > theta ( k )(x,y)-model (66.0/66.7). We also carried out a 2D-E analysis of biological samples of L. infantum promastigotes focusing on a 2D-E gel spot of one unknown protein with M<20, 100 and pI <7. MASCOT search identified 20 proteins with Mowse score >30, but not one >52 (threshold value), the higher value of 42 was for a probable DNA-directed RNA polymerase. However, we determined experimentally the sequence of more than 140 peptides. We used QSAR models to predict RNase scores for these peptides and BLAST alignment to confirm some results. We also calculated 3D-folding TIs based on MD experiments and compared 2D versus 3D-TIs on molecular phylogenetic analysis of the molecular diversity of these peptides. This combined strategy may be of interest in drug development or target identification.

Annotation-based inference of transporter function

Motivation: We present a method for inferring and constructing transport reactions for transporter proteins based primarily on the analysis of the names of individual proteins in the genome annotation of an organism. Transport reactions are declarative descriptions of transporter activities, and thus can be manipulated computationally, unlike free-text protein names. Once transporter activities are encoded as transport reactions, a number of computational analyses are possible including database queries by transporter activity; inclusion of transporters into an automatically generated metabolic-map diagram that can be painted with omics data to aid in their interpretation; detection of anomalies in the metabolic and transport networks, such as substrates that are transported into the cell but are not inputs to any metabolic reaction or pathway; and comparative analyses of the transport capabilities of different organisms.

Results: On randomly selected organisms, the method achieves precision and recall rates of 0.93 and 0.90, respectively in identifying transporter proteins by name within the complete genome. The method obtains 67.5% accuracy in predicting complete transport reactions; if allowance is made for predictions that are overly general yet not incorrect, reaction prediction accuracy is 82.5%.

Availability: The method is implemented as part of PathoLogic, the inference component of the Pathway Tools software. Pathway Tools is freely available to researchers at non-commercial institutions, including source code; a fee applies to commercial institutions.

Contact:tomlee@ai.sri.com

Supplementary information:Supplementary data are available at Bioinformatics online.

Prediction of antibiotic resistance proteins from sequence-derived properties irrespective of sequence similarity

Increasing antibiotic resistance has become a worldwide challenge to the clinical treatment of infectious diseases. The identification of antibiotic resistance proteins (ARPs) would be helpful in the discovery of new therapeutic targets and the design of novel drugs to control the potential spread of antibiotic resistance. In this work, a support vector machine (SVM)-based ARP prediction system was developed using 1308 ARPs and 15587 non-ARPs. Its performance was evaluated using 313 ARPs and 7156 non-ARPs. The computed prediction accuracy was 88.5% for ARPs and 99.2% for non-ARPs. A potential application of this method is the identification of ARPs non-homologous to proteins of known function. Further genome screening found that ca. 3.5% and 3.2% of proteins in Escherichia coli and Staphylococcus aureus, respectively, are potential ARPs. These results suggest the usefulness of SVMs for facilitating the identification of ARPs. The software can be accessed at SARPI (Server for Antibiotic Resistance Protein Identification).

Using SELDI-TOF MS to identify serum biomarkers of rheumatoid arthritis

OBJECTIVES

No satisfactory biomarkers are currently available to screen for rheumatoid arthritis (RA). We have developed and evaluated surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) for detection and analysis of multiple proteins for distinguishing individuals with RA from control individuals.

METHODS

A total of 156 serum samples from 90 RA patients, 30 patients with ankylosing spondylitis (AS), and 36 healthy individuals were examined by SELDI technology. Spectral data were analysed by the support vector machine (SVM) approach and potential biomarkers were chosen for system training and were used to construct a diagnostic model.

RESULTS

Pattern 1, consisting of four protein peaks with m/z values of 3899, 4594, 7566, and 13,842, distinguished RA from the healthy samples with sensitivity of 90.0% and a specificity of 91.7%. Pattern 2, consisting of m/z peaks 4287 and 6471, distinguished RA from AS with a sensitivity of 86.7% and a specificity of 85.0%.

CONCLUSION

The combination of SELDI-TOF MS and SVM could facilitate the discovery of better biomarkers for RA and also provide a useful tool for molecular diagnosis in the future.

Evaluation of virtual screening performance of support vector machines trained by sparsely distributed active compounds

Virtual screening performance of support vector machines (SVM) depends on the diversity of training active and inactive compounds. While diverse inactive compounds can be routinely generated, the number and diversity of known actives are typically low. We evaluated the performance of SVM trained by sparsely distributed actives in six MDDR biological target classes composed of a high number of known actives (983-1645) of high, intermediate, and low structural diversity (muscarinic M1 receptor agonists, NMDA receptor antagonists, thrombin inhibitors, HIV protease inhibitors, cephalosporins, and renin inhibitors). SVM trained by regularly sparse data sets of 100 actives show improved yields at substantially reduced false-hit rates compared to those of published studies and those of Tanimoto-based similarity searching method based on the same data sets and molecular descriptors. SVM trained by very sparse data sets of 40 actives (2.4%-4.1% of the known actives) predicted 17.5-39.5%, 23.0-48.1%, and 70.2-92.4% of the remaining 943-1605 actives in the high, intermediate, and low diversity classes, respectively, 13.8-68.7% of which are outside the training compound families. SVM predicted 99.97% and 97.1% of the 9.997 M PUBCHEM and 167K remaining MDDR compounds as inactive and 2.6%-8.3% of the 19,495-38,483 MDDR compounds similar to the known actives as active. These suggest that SVM has substantial capability in identifying novel active compounds from sparse active data sets at low false-hit rates.

Investigation of transmembrane proteins using a computational approach

BACKGROUND

An important subfamily of membrane proteins are the transmembrane alpha-helical proteins, in which the membrane-spanning regions are made up of alpha-helices. Given the obvious biological and medical significance of these proteins, it is of tremendous practical importance to identify the location of transmembrane segments. The difficulty of inferring the secondary or tertiary structure of transmembrane proteins using experimental techniques has led to a surge of interest in applying techniques from machine learning and bioinformatics to infer secondary structure from primary structure in these proteins. We are therefore interested in determining which physicochemical properties are most useful for discriminating transmembrane segments from non-transmembrane segments in transmembrane proteins, and for discriminating intrinsically unstructured segments from intrinsically structured segments in transmembrane proteins, and in using the results of these investigations to develop classifiers to identify transmembrane segments in transmembrane proteins.

RESULTS

We determined that the most useful properties for discriminating transmembrane segments from non-transmembrane segments and for discriminating intrinsically unstructured segments from intrinsically structured segments in transmembrane proteins were hydropathy, polarity, and flexibility, and used the results of this analysis to construct classifiers to discriminate transmembrane segments from non-transmembrane segments using four classification techniques: two variants of the Self-Organizing Global Ranking algorithm, a decision tree algorithm, and a support vector machine algorithm. All four techniques exhibited good performance, with out-of-sample accuracies of approximately 75%.

CONCLUSIONS

Several interesting observations emerged from our study: intrinsically unstructured segments and transmembrane segments tend to have opposite properties; transmembrane proteins appear to be much richer in intrinsically unstructured segments than other proteins; and, in approximately 70% of transmembrane proteins that contain intrinsically unstructured segments, the intrinsically unstructured segments are close to transmembrane segments.

Detection and significance of serum protein markers of small-cell lung cancer

Currently, no satisfactory biomarkers are available to screen for small-cell lung cancer (SCLC). We applied a surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) ProteinChip system to detect 150 serum samples (including 54 SCLC patients, 24 non-small cell lung cancer [NSCLC] patients, 32 pneumonia patients, and 40 healthy individuals). The spectra data were analyzed by support vector machine (SVM) and potential biomarkers were chosen for the system training and used to construct diagnostic model. Pattern 1, constructed of four protein peaks with mass/charge (m/z) of 4,293 Da, 4,612 Da, 6,455 Da, and 7,582 Da, separated SCLC patients from the healthy individuals with a sensitivity of 88.9% and a specificity of 85.7%. This pattern performed significantly better than the current marker, neuron-specific enolase (NSE) (P<0.05). Pattern 2, constructed of protein peaks with mass/charge (m/z) of 2,764 Da and 1,7368 Da, separated SCLC from pneumonia with a sensitivity of 88.9% and a specificity of 91.7%. Pattern 3, constructed of another three protein peaks with m/z of 3,912 Da, 7,562 Da, and 13,777 Da, separated SCLC from NSCLC. The sensitivity and specificity were 83.3% and 75.0%, respectively. These results suggested that SELDI-TOF MS combined with support vector machine yields significantly higher sensitivity and specificity for the detection of serum protein of SCLC.

A support vector machines approach for virtual screening of active compounds of single and multiple mechanisms from large libraries at an improved hit-rate and enrichment factor

Support vector machines (SVM) and other machine-learning (ML) methods have been explored as ligand-based virtual screening (VS) tools for facilitating lead discovery. While exhibiting good hit selection performance, in screening large compound libraries, these methods tend to produce lower hit-rate than those of the best performing VS tools, partly because their training-sets contain limited spectrum of inactive compounds. We tested whether the performance of SVM can be improved by using training-sets of diverse inactive compounds. In retrospective database screening of active compounds of single mechanism (HIV protease inhibitors, DHFR inhibitors, dopamine antagonists) and multiple mechanisms (CNS active agents) from large libraries of 2.986 million compounds, the yields, hit-rates, and enrichment factors of our SVM models are 52.4-78.0%, 4.7-73.8%, and 214-10,543, respectively, compared to those of 62-95%, 0.65-35%, and 20-1200 by structure-based VS and 55-81%, 0.2-0.7%, and 110-795 by other ligand-based VS tools in screening libraries of >or=1 million compounds. The hit-rates are comparable and the enrichment factors are substantially better than the best results of other VS tools. 24.3-87.6% of the predicted hits are outside the known hit families. SVM appears to be potentially useful for facilitating lead discovery in VS of large compound libraries.

Efficacy of different protein descriptors in predicting protein functional families

Background

Sequence-derived structural and physicochemical descriptors have frequently been used in machine learning prediction of protein functional families, thus there is a need to comparatively evaluate the effectiveness of these descriptor-sets by using the same method and parameter optimization algorithm, and to examine whether the combined use of these descriptor-sets help to improve predictive performance. Six individual descriptor-sets and four combination-sets were evaluated in support vector machines (SVM) prediction of six protein functional families.

Results

The performance of these descriptor-sets were ranked by Matthews correlation coefficient (MCC), and categorized into two groups based on their performance. While there is no overwhelmingly favourable choice of descriptor-sets, certain trends were found. The combination-sets tend to give slightly but consistently higher MCC values and thus overall best performance such that three out of four combination-sets show slightly better performance compared to one out of six individual descriptor-sets.

Conclusion

Our study suggests that currently used descriptor-sets are generally useful for classifying proteins and the prediction performance may be enhanced by exploring combinations of descriptors.

KinasePhos 2.0: a web server for identifying protein kinase-specific phosphorylation sites based on sequences and coupling patterns

Due to the importance of protein phosphorylation in cellular control, many researches are undertaken to predict the kinase-specific phosphorylation sites. Referred to our previous work, KinasePhos 1.0, incorporated profile hidden Markov model (HMM) with flanking residues of the kinase-specific phosphorylation sites. Herein, a new web server, KinasePhos 2.0, incorporates support vector machines (SVM) with the protein sequence profile and protein coupling pattern, which is a novel feature used for identifying phosphorylation sites. The coupling pattern [XdZ] denotes the amino acid coupling-pattern of amino acid types X and Z that are separated by d amino acids. The differences or quotients of coupling strength C_XdZ between the positive set of phosphorylation sites and the background set of whole protein sequences from Swiss-Prot are computed to determine the number of coupling patterns for training SVM models. After the evaluation based on k-fold cross-validation and Jackknife cross-validation, the average predictive accuracy of phosphorylated serine, threonine, tyrosine and histidine are 90, 93, 88 and 93%, respectively. KinasePhos 2.0 performs better than other tools previously developed. The proposed web server is freely available at http://KinasePhos2.mbc.nctu.edu.tw/.

RF-DYMHC: detecting the yeast meiotic recombination hotspots and coldspots by random forest model using gapped dinucleotide composition features

In the yeast, meiotic recombination is initiated by double-strand DNA breaks (DSBs) which occur at relatively high frequencies in some genomic regions (hotspots) and relatively low frequencies in others (coldspots). Although observations concerning individual hot/cold spots have given clues as to the mechanism of recombination initiation, the prediction of hot/cold spots from DNA sequence information is a challenging task. In this article, we introduce a random forest (RF) prediction model to detect recombination hot/cold spots from yeast genome. The out-of-bag (OOB) estimation of the model indicated that the RF classifier achieved high prediction performance with 82.05% total accuracy and 0.638 Mattew's correlation coefficient (MCC) value. Compared with an alternative machine-learning algorithm, support vector machine (SVM), the RF method outperforms it in both sensitivity and specificity. The prediction model is implemented as a web server (RF-DYMHC) and it is freely available at http://www.bioinf.seu.edu.cn/Recombination/rf_dymhc.htm. Given a yeast genome and prediction parameters (RI-value and non-overlapping window scan size), the program reports the predicted hot/cold spots and marks them in color.

Support vector machines approach for predicting druggable proteins: recent progress in its exploration and investigation of its usefulness

Identification and validation of viable targets is an important first step in drug discovery and new methods, and integrated approaches are continuously explored to improve the discovery rate and exploration of new drug targets. An in silico machine learning method, support vector machines, has been explored as a new method for predicting druggable proteins from amino acid sequence independent of sequence similarity, thereby facilitating the prediction of druggable proteins that exhibit no or low homology to known targets.

Prediction of the functional class of metal-binding proteins from sequence derived physicochemical properties by support vector machine approach

Metal-binding proteins play important roles in structural stability, signaling, regulation, transport, immune response, metabolism control, and metal homeostasis. Because of their functional and sequence diversity, it is desirable to explore additional methods for predicting metal-binding proteins irrespective of sequence similarity. This work explores support vector machines (SVM) as such a method. SVM prediction systems were developed by using 53,333 metal-binding and 147,347 non-metal-binding proteins, and evaluated by an independent set of 31,448 metal-binding and 79,051 non-metal-binding proteins. The computed prediction accuracy is 86.3%, 81.6%, 83.5%, 94.0%, 81.2%, 85.4%, 77.6%, 90.4%, 90.9%, 74.9% and 78.1% for calcium-binding, cobalt-binding, copper-binding, iron-binding, magnesium-binding, manganese-binding, nickel-binding, potassium-binding, sodium-binding, zinc-binding, and all metal-binding proteins respectively. The accuracy for the non-member proteins of each class is 88.2%, 99.9%, 98.1%, 91.4%, 87.9%, 94.5%, 99.2%, 99.9%, 99.9%, 98.0%, and 88.0% respectively. Comparable accuracies were obtained by using a different SVM kernel function. Our method predicts 67% of the 87 metal-binding proteins non-homologous to any protein in the Swissprot database and 85.3% of the 333 proteins of known metal-binding domains as metal-binding. These suggest the usefulness of SVM for facilitating the prediction of metal-binding proteins. Our software can be accessed at the SVMProt server http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi.

Recent progresses in the application of machine learning approach for predicting protein functional class independent of sequence similarity

Protein sequence contains clues to its function. Functional prediction from sequence presents a challenge particularly for proteins that have low or no sequence similarity to proteins of known function. Recently, machine learning methods have been explored for predicting functional class of proteins from sequence-derived properties independent of sequence similarity, which showed promising potential for low- and non-homologous proteins. These methods can thus be explored as potential tools to complement alignment- and clustering-based methods for predicting protein function. This article reviews the strategies, current progresses, and underlying difficulties in using machine learning methods for predicting the functional class of proteins. The relevant software and web-servers are described. The reported prediction performances in the application of these methods are also presented, which need to be interpreted with caution as they are dependent on such factors as datasets used and choice of parameters.

PROFEAT: a web server for computing structural and physicochemical features of proteins and peptides from amino acid sequence

Sequence-derived structural and physicochemical features have frequently been used in the development of statistical learning models for predicting proteins and peptides of different structural, functional and interaction profiles. PROFEAT (Protein Features) is a web server for computing commonly-used structural and physicochemical features of proteins and peptides from amino acid sequence. It computes six feature groups composed of ten features that include 51 descriptors and 1447 descriptor values. The computed features include amino acid composition, dipeptide composition, normalized Moreau-Broto autocorrelation, Moran autocorrelation, Geary autocorrelation, sequence-order-coupling number, quasi-sequence-order descriptors and the composition, transition and distribution of various structural and physicochemical properties. In addition, it can also compute previous autocorrelations descriptors based on user-defined properties. Our computational algorithms were extensively tested and the computed protein features have been used in a number of published works for predicting proteins of functional classes, protein-protein interactions and MHC-binding peptides. PROFEAT is accessible at http://jing.cz3.nus.edu.sg/cgi-bin/prof/prof.cgi.

Prediction of the functional class of lipid binding proteins from sequence-derived properties irrespective of sequence similarity

Lipid binding proteins play important roles in signaling, regulation, membrane trafficking, immune response, lipid metabolism, and transport. Because of their functional and sequence diversity, it is desirable to explore additional methods for predicting lipid binding proteins irrespective of sequence similarity. This work explores the use of support vector machines (SVMs) as such a method. SVM prediction systems are developed using 14,776 lipid binding and 133,441 nonlipid binding proteins and are evaluated by an independent set of 6,768 lipid binding and 64,761 nonlipid binding proteins. The computed prediction accuracy is 78.9, 79.5, 82.2, 79.5, 84.4, 76.6, 90.6, 79.0, and 89.9% for lipid degradation, lipid metabolism, lipid synthesis, lipid transport, lipid binding, lipopolysaccharide biosynthesis, lipoprotein, lipoyl, and all lipid binding proteins, respectively. The accuracy for the nonmember proteins of each class is 99.9, 99.2, 99.6, 99.8, 99.9, 99.8, 98.5, 99.9, and 97.0%, respectively. Comparable accuracies are obtained when homologous proteins are considered as one, or by using a different SVM kernel function. Our method predicts 86.8% of the 76 lipid binding proteins nonhomologous to any protein in the Swiss-Prot database and 89.0% of the 73 known lipid binding domains as lipid binding. These findings suggest the usefulness of SVMs for facilitating the prediction of lipid binding proteins. Our software can be accessed at the SVMProt server (http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi).