Reduced bio-basis function neural networks for protease cleavage site prediction

This paper presents a new neural learning algorithm for protease cleavage site prediction. The basic idea is to replace the radial basis function used in radial basis function neural networks by a so-called bio-basis function using amino acid similarity matrices. Mutual information is used to select bio-bases and a corresponding selection algorithm is developed. The algorithm has been applied to the prediction of HIV and Hepatitis C virus protease cleavage sites in proteins with success.     

Predicting the linkage sites in glycoproteins using bio-basis function neural network

MOTIVATION: Although, it is known that O-glycosidically linked oligosaccharides are commonly conjugated to a serine, threonine or hydroxylysine residue of the polypeptide, the chemical nature of the anchoring monosaccharide and the size of the oligosaccharide unit varies. Among different types, O-linked or mucin-type oligosaccharides are intimately involved in the secretion of proteins, be they enzymes, hormones or structural glycoproteins. Knowledge of the linkage sites in glycoproteins is critical to the design of specific and efficient inhibitors against the enzyme to catalyse the formation of the carbohydrate-peptide linkage. RESULTS: We present a method for predicting the linkage sites in O-linked glycoproteins using bio-basis function neural networks. The mean prediction accuracy of this method is 91.15 +/- 2.75% while it is 82.28 +/- 6.45% using back-propagation neural networks. Importantly, this method has significantly reduced the CPU time for modelling.     

Prediction of Signal Peptides Using Bio-Basis Function Neural Networks and Decision Trees

Signal peptide identification is of immense importance in drug design. Accurate identification of signal peptides is the first critical step to be able to change the direction of the targeting proteins and use the designed drug to target a specific organelle to correct a defect. Because experimental identification is the most accurate method, but is expensive and time-consuming, an efficient and affordable automated system is of great interest. In this article, we propose using an adapted neural network, called a bio-basis function neural network, and decision trees for predicting signal peptides. The bio-basis function neural network model and decision trees achieved 97.16% and 97.63% accuracy respectively, demonstrating that the methods work well for the prediction of signal peptides. Moreover, decision trees revealed that position P(1'), which is important in forming signal peptides, most commonly comprises either leucine or alanine. This concurs with the (P(3)-P(1)-P(1')) coupling model.     

CASVM: web server for SVM-based prediction of caspase substrates cleavage sites

Caspases belong to a unique class of cysteine proteases which function as critical effectors of apoptosis, inflammation and other important cellular processes. Caspases cleave substrates at specific tetrapeptide sites after a highly conserved aspartic acid residue. Prediction of such cleavage sites will complement structural and functional studies on substrates cleavage as well as discovery of new substrates. We have recently developed a support vector machines (SVM) method to address this issue. Our algorithm achieved an accuracy ranging from 81.25 to 97.92%, making it one of the best methods currently available. CASVM is the web server implementation of our SVM algorithms, written in Perl and hosted on a Linux platform. The server can be used for predicting non-canonical caspase substrate cleavage sites. We have also included a relational database containing experimentally verified caspase substrates retrievable using accession IDs, keywords or sequence similarity. AVAILABILITY: http://www.casbase.org/casvm/index.html     

Adaptive encoding neural networks for the recognition of human signal peptide cleavage sites

MOTIVATION: Data representation and encoding are essential for classification of protein sequences with artificial neural networks (ANN). Biophysical properties are appropriate for low dimensional encoding of protein sequence data. However, in general there is no a priori knowledge of the relevant properties for extraction of representative features. RESULTS: An adaptive encoding artificial neural network (ACN) for recognition of sequence patterns is described. In this approach parameters for sequence encoding are optimized within the same process as the weight vectors by an evolutionary algorithm. The method is applied to the prediction of signal peptide cleavage sites in human secretory proteins and compared with an established predictor for signal peptides. CONCLUSION: Knowledge of physico-chemical properties is not necessary for training an ACN. The advantage is a low dimensional data representation leading to computational efficiency, easy evaluation of the detected features, and high prediction accuracy. Availability: A cleavage site prediction server is located at the Humboldt University http://itb.biologie.hu-berlin.de/ approximately jo/sig-cleave/ACNpredictor.cgi Contact: jo@itb.hu-berlin.de; berndj@zedat.fu-berlin.de     

Cleavage site analysis in picornaviral polyproteins: discovering cellular targets by neural networks.

Picornaviral proteinases are responsible for maturation cleavages of the viral polyprotein, but also catalyze the degradation of cellular targets. Using graphical visualization techniques and neural network algorithms, we have investigated the sequence specificity of the two proteinases 2Apro and 3Cpro. The cleavage of VP0 (giving rise to VP2 and VP4), which is carried out by a so-far unknown proteinase, was also examined. In combination with a novel surface exposure prediction algorithm, our neural network approach successfully distinguishes known cleavage sites from noncleavage sites and yields a more consistent definition of features common to these sites. The method is able to predict experimentally determined cleavage sites in cellular proteins. We present a list of mammalian and other proteins that are predicted to be possible targets for the viral proteinases. Whether these proteins are indeed cleaved awaits experimental verification. Additionally, we report several errors detected in the protein databases. A computer server for prediction of cleavage sites by picornaviral proteinases is publicly available at the e-mail address NetPicoRNA@cbs.dtu.dk or via WWW at http:@www.cbs.dtu.dk/services/NetPicoRNA/.     

Predicting protein secondary structure using neural net and statistical methods.

A comparison of neural network methods and Bayesian statistical methods is presented for prediction of the secondary structure of proteins given their primary sequence. The Bayesian method makes the unphysical assumption that the probability of an amino acid occurring in each position in the protein is independent of the amino acids occurring elsewhere. However, we find the predictive accuracy of the Bayesian method to be only minimally less than the accuracy of the most sophisticated methods used to date. We present the relationship of neural network methods to Bayesian statistical methods and show that, in principle, neural methods offer considerable power, although apparently they are not particularly useful for this problem. In the process, we derive a neural formalism in which the output neurons directly represent the conditional probabilities of structure class. The probabilistic formalism allows introduction of a new objective function, the mutual information, which translates the notion of correlation as a measure of predictive accuracy into a useful training measure. Although a similar accuracy to other approaches (utilizing a mean-square error) is achieved using this new measure, the accuracy on the training set is significantly and tantalizingly higher, even though the number of adjustable parameters remains the same. The mutual information measure predicts a greater fraction of helix and sheet structures correctly than the mean-square error measure, at the expense of coil accuracy, precisely as it was designed to do. By combining the two objective functions, we obtain a marginally improved accuracy of 64.4%, with Matthews coefficients C alpha, C beta and Ccoil of 0.40, 0.32 and 0.42, respectively. However, since all methods to date perform only slightly better than the Bayes algorithm, which entails the drastic assumption of independence of amino acids, one is forced to conclude that little progress has been made on this problem, despite the application of a variety of sophisticated algorithms such as neural networks, and that further advances will require a better understanding of the relevant biophysics.     

Global sequencing of proteolytic cleavage sites in apoptosis by specific labeling of protein N termini

The nearly 600 proteases in the human genome regulate a diversity of biological processes, including programmed cell death. Comprehensive characterization of protease signaling in complex biological samples is limited by available proteomic methods. We have developed a general approach for global identification of proteolytic cleavage sites using an engineered enzyme to selectively biotinylate free protein N termini for positive enrichment of corresponding N-terminal peptides. Using this method to study apoptosis, we have sequenced 333 caspase-like cleavage sites distributed among 292 protein substrates. These sites are generally not predicted by in vitro caspase substrate specificity but can be used to predict other physiological caspase cleavage sites. Structural bioinformatic studies show that caspase cleavage sites often appear in surface-accessible loops and even occasionally in helical regions. Strikingly, we also find that a disproportionate number of caspase substrates physically interact, suggesting that these dimeric proteases target protein complexes and networks to elicit apoptosis.     

Identification of Caspase Cleavage Sites in KSHV Latency-Associated Nuclear Antigen and Their Effects on Caspase-Related Host Defense Responses

Kaposi’s sarcoma-associated herpesvirus (KSHV), also known as human herpesvirus-8, is the causative agent of three hyperproliferative disorders: Kaposi’s sarcoma, primary effusion lymphoma (PEL) and multicentric Castleman’s disease. During viral latency a small subset of viral genes are produced, including KSHV latency-associated nuclear antigen (LANA), which help the virus thwart cellular defense responses. We found that exposure of KSHV-infected cells to oxidative stress, or other inducers of apoptosis and caspase activation, led to processing of LANA and that this processing could be inhibited with the pan-caspase inhibitor Z-VAD-FMK. Using sequence, peptide, and mutational analysis, two caspase cleavage sites within LANA were identified: a site for caspase-3 type caspases at the N-terminus and a site for caspase-1 and-3 type caspases at the C-terminus. Using LANA expression plasmids, we demonstrated that mutation of these cleavage sites prevents caspase-1 and caspase-3 processing of LANA. This indicates that these are the principal sites that are susceptible to caspase cleavage. Using peptides spanning the identified LANA cleavage sites, we show that caspase activity can be inhibited in vitro and that a cell-permeable peptide spanning the C-terminal cleavage site could inhibit cleavage of poly (ADP-ribose) polymerase and increase viability in cells undergoing etoposide-induced apoptosis. The C-terminal peptide of LANA also inhibited interleukin-1beta (IL-1β) production from lipopolysaccharide-treated THP-1 cells by more than 50%. Furthermore, mutation of the two cleavage sites in LANA led to a significant increase in IL-1β production in transfected THP-1 cells; this provides evidence that these sites function to blunt the inflammasome, which is known to be activated in latently infected PEL cells. These results suggest that specific caspase cleavage sites in KSHV LANA function to blunt apoptosis as well as interfere with the caspase-1-mediated inflammasome, thus thwarting key cellular defense mechanisms.     

Feature-extraction from endopeptidase cleavage sites in mitochondrial targeting peptides

Cleavage sites in nuclear-encoded mitochondrial protein targeting peptides (mTPs) from mammals, yeast, and plants have been analysed for characteristic physicochemical features using statistical methods, perceptrons, multilayer neural networks, and self-organizing feature maps. Three different sequence motifs were found, revealing loosely defined arginine motifs with Arg in positions −10, −3, and −2. A self-organizing feature map was able to cluster these three types of endopeptidase target sites but did not identify any species-specific characteristics in mTPs. Neural networks were used to define local sequence features around precursor cleavage sites. Proteins 30:49–60, 1998. © 1998 Wiley-Liss, Inc.     

Regulation of cell proliferation and survival: Convergence of protein kinases and caspases

Intricate networks of protein kinases are intimately involved in the regulation of cellular events related to cell proliferation and survival. In addition to protein kinases, cells also contain networks of proteases including aspartic-acid directed caspases organized in cascades that play a major role in the regulation of cell survival through their involvement in the initiation and execution phases of apoptosis. Perturbations in regulatory protein kinase and caspase networks induce alterations in cell survival and frequently accompany transformation and tumorigenesis. Furthermore, recent studies have documented that caspases or their substrates are subject to phosphorylation in cells illustrating a potential convergence of protein kinase and caspase signaling pathways. Interestingly, a number of caspase substrates are protected from cleavage when they are phosphorylated at sites that are adjacent to caspase cleavage sites. While it is theoretically possible that many distinct protein kinases could protect proteins from caspase-mediated cleavage, protein kinase CK2 is of particular interest because acidic amino acids, including aspartic acid residues that are recognized by caspases, are its dominant specificity determinants.     

Cross-validation of protein structural class prediction using statistical clustering and neural networks.

We present an approach to predicting protein structural class that uses amino acid composition and hydrophobic pattern frequency information as input to two types of neural networks: (1) a three-layer back-propagation network and (2) a learning vector quantization network. The results of these methods are compared to those obtained from a modified Euclidean statistical clustering algorithm. The protein sequence data used to drive these algorithms consist of the normalized frequency of up to 20 amino acid types and six hydrophobic amino acid patterns. From these frequency values the structural class predictions for each protein (all-alpha, all-beta, or alpha-beta classes) are derived. Examples consisting of 64 previously classified proteins were randomly divided into multiple training (56 proteins) and test (8 proteins) sets. The best performing algorithm on the test sets was the learning vector quantization network using 17 inputs, obtaining a prediction accuracy of 80.2%. The Matthews correlation coefficients are statistically significant for all algorithms and all structural classes. The differences between algorithms are in general not statistically significant. These results show that information exists in protein primary sequences that is easily obtainable and useful for the prediction of protein structural class by neural networks as well as by standard statistical clustering algorithms.     

Protein C‐terminal enzymatic labeling identifies novel caspase cleavages during the apoptosis of multiple myeloma cells induced by kinase inhibition

Caspase activation and proteolytic cleavages are the major events in the early stage of apoptosis. Identification of protein substrates cleaved by caspases will reveal the occurrence of the early events in the apoptotic process and may provide potential drug targets for cancer therapy. Although several N‐terminal MS‐based proteomic approaches have been developed to identify proteolytic cleavages, these methods have their inherent drawbacks. Here we apply a previously developed proteomic approach, protein C‐terminal enzymatic labeling (ProC‐TEL), to identify caspase cleavage events occurring in the early stage of the apoptosis of a myeloma cell line induced by kinase inhibition. Both previously identified and novel caspase cleavage sites are detected and the reduction of the expression level of several proteins is confirmed biochemically upon kinase inhibition although the current ProC‐TEL procedure is not fully optimized to provide peptide identifications comparable to N‐terminal labeling approaches. The identified cleaved proteins form a complex interaction network with central hubs determining morphological changes during the apoptosis. Sequence analyses show that some ProC‐TEL identified caspase cleavage events are unidentifiable when traditional N‐terminomic approaches are utilized. This work demonstrates that ProC‐TEL is a complementary approach to the N‐terminomics for the identification of proteolytic cleavage events such as caspase cleavages in signaling pathways.     

Comparative analysis of neuropeptide cleavage sites in human, mouse, rat, and cattle

Neuropeptides are an important class of signaling molecules that result from complex and variable posttranslational processing of precursor proteins and thus are difficult to identify based solely on genomic information. Bioinformatics prediction of precursor cleavage sites can support effective biochemical characterization of neuropeptides. Neuropeptide cleavage models were developed using comprehensive human, mouse, rat, and cattle precursor data sets and used to compare predicted neuropeptide processing across these species. Logistic regression and artificial neural network models were used to predict cleavages based on amino acid and physiochemical properties of amino acids at precursor sequence locations proximal to cleavage. Correct cleavage classification rates across species and models ranged from 85% to 100%, suggesting that amino acid and amino acid properties have major impact on the probability of cleavage and that these factors have comparable effects in human, mouse, rat, and cattle. The variable accuracy of each species-specific model to predict cleavage sites indicated that there are species- and precursor-specific processing patterns. Prediction of mouse cleavages using rat models was highly accurate, yet the reverse was not observed. Sensitivity and specificity revealed that logistic models are well suited to maximize the rate of true noncleavage predictions with moderate rates of true cleavage predictions; meanwhile, artificial neural networks maximize the rate of true cleavage predictions with moderate to low true noncleavage predictions. Logistic models also provided insights into the strength of the amino acid associations with cleavage. Prediction of neuropeptide cleavage sites using human, mouse, rat, and cattle models are available at . Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users. Allison Tegge and Bruce Southey contributed equally to this work.     

Identification of ribosome binding sites in Escherichia coli using neural network models.

This study investigated the use of neural networks in the identification of Escherichia coli ribosome binding sites. The recognition of these sites based on primary sequence data is difficult due to the multiple determinants that define them. Additionally, secondary structure plays a significant role in the determination of the site and this information is difficult to include in the models. Efforts to solve this problem have so far yielded poor results. A new compilation of E. coli ribosome binding sites was generated for this study. Feedforward backpropagation networks were applied to their identification. Perceptrons were also applied, since they have been the previous best method since 1982. Evaluation of performance for all the neural networks and perceptrons was determined by ROC analysis. The neural network provided significant improvement in the recognition of these sites when compared with the previous best method, finding less than half the number of false positives when both models were adjusted to find an equal number of actual sites. The best neural network used an input window of 101 nucleotides and a single hidden layer of 9 units. Both the neural network and the perceptron trained on the new compilation performed better than the original perceptron published by Stormo et al. in 1982.     

Caspase Cleavage Sites in the Human Proteome: CaspDB,a Database of Predicted Substrates

Caspases are enzymes belonging to a conserved family of cysteine-dependent aspartic-specific proteases that are involved in vital cellular processes and play a prominent role in apoptosis and inflammation. Determining all relevant protein substrates of caspases remains a challenging task. Over 1500 caspase substrates have been discovered in the human proteome according to published data and new substrates are discovered on a daily basis. To aid the discovery process we developed a caspase cleavage prediction method using the recently published curated MerCASBA database of experimentally determined caspase substrates and a Random Forest classification method. On both internal and external test sets, the ranking of predicted cleavage positions is superior to all previously developed prediction methods. The in silico predicted caspase cleavage positions in human proteins are available from a relational database: CaspDB. Our database provides information about potential cleavage sites in a verified set of all human proteins collected in Uniprot and their orthologs, allowing for tracing of cleavage motif conservation. It also provides information about the positions of disease-annotated single nucleotide polymorphisms, and posttranslational modifications that may modulate the caspase cleaving efficiency.     

Caspase cleavage of the nonstructural protein NS1 mediates replication of Aleutian mink disease parvovirus

Virus-induced apoptosis of infected cells can limit both the time and the cellular machinery available for virus replication. Hence, many viruses have evolved strategies to specifically inhibit apoptosis. However, Aleutian mink disease parvovirus (ADV) is the first example of a DNA virus that not only induces apoptosis but also utilizes caspase activity to facilitate virus replication. To determine the function of caspase activity during ADV replication, virus-infected cell lysates or purified ADV proteins were incubated with various purified caspases. Caspases cleaved the major nonstructural protein of ADV (NS1) at two caspase recognition sequences, whereas ADV structural proteins could not be cleaved. Importantly, the NS1 products could be identified in ADV-infected cells but were not present in infected cells pretreated with caspase inhibitors. By mutating putative caspase cleavage sites (D to E), we mapped the two cleavage sites to amino acid residues NS1:227 (INTD downward arrow S) and NS1:285 (DQTD downward arrow S). Replication of ADV containing either of these mutations was reduced 10(3)- to 10(4)-fold compared to that of wild-type virus, and a construct containing both mutations was replication defective. Immunofluorescent studies revealed that cleavage was required for nuclear localization of NS1. The requirement for caspase activity during permissive replication suggests that limitation of caspase activation and apoptosis in vivo may be a novel approach to restricting virus replication.     

Improved prediction of signal peptides: SignalP 3.0

We describe improvements of the currently most popular method for prediction of classically secreted proteins, SignalP. SignalP consists of two different predictors based on neural network and hidden Markov model algorithms, where both components have been updated. Motivated by the idea that the cleavage site position and the amino acid composition of the signal peptide are correlated, new features have been included as input to the neural network. This addition, combined with a thorough error-correction of a new data set, have improved the performance of the predictor significantly over SignalP version 2. In version 3, correctness of the cleavage site predictions has increased notably for all three organism groups, eukaryotes, Gram-negative and Gram-positive bacteria. The accuracy of cleavage site prediction has increased in the range 6-17% over the previous version, whereas the signal peptide discrimination improvement is mainly due to the elimination of false-positive predictions, as well as the introduction of a new discrimination score for the neural network. The new method has been benchmarked against other available methods. Predictions can be made at the publicly available web server