ATGPred-FL: sequence-based prediction of autophagy proteins with feature representation learning

Amino Acids - Autophagy plays an important role in biological evolution and is regulated by many autophagy proteins. Accurate identification of autophagy proteins is crucially important to reveal...     

Drug response prediction by ensemble learning and drug-induced gene expression signatures

Chemotherapeutic response of cancer cells to a given compound is one of the most fundamental information one requires to design anti-cancer drugs. Recently, considerable amount of drug-induced gene expression data has become publicly available, in addition to cytotoxicity databases. These large sets of data provided an opportunity to apply machine learning methods to predict drug activity. However, due to the complexity of cancer drug mechanisms, none of the existing methods is perfect. In this paper, we propose a novel ensemble learning method to predict drug response. In addition, we attempt to use the drug screen data together with two novel signatures produced from the drug-induced gene expression profiles of cancer cell lines. Finally, we evaluate predictions by in vitro experiments in addition to the tests on data sets. The predictions of the methods, the signatures and the software are available from http://mtan.etu.edu.tr/drug-response-prediction/.     

Choosing SNPs using feature selection

A major challenge for genomewide disease association studies is the high cost of genotyping large number of single nucleotide polymorphisms (SNPs). The correlations between SNPs, however, make it possible to select a parsimonious set of informative SNPs, known as "tagging" SNPs, able to capture most variation in a population. Considerable research interest has recently focused on the development of methods for finding such SNPs. In this paper, we present an efficient method for finding tagging SNPs. The method does not involve computation-intensive search for SNP subsets but discards redundant SNPs using a feature selection algorithm. In contrast to most existing methods, the method presented here does not limit itself to using only correlations between SNPs in local groups. By using correlations that occur across different chromosomal regions, the method can reduce the number of globally redundant SNPs. Experimental results show that the number of tagging SNPs selected by our method is smaller than by using block-based methods. Supplementary website: http://htsnp.stanford.edu/FSFS/.     

GalNAc-transferase specificity prediction based on feature selection method

GalNAc-transferase can catalyze the biosynthesis of O-linked oligosaccharides. The specificity of GalNAc-transferase is composed of nine amino acid residues denoted by R4, R3, R2, R1, R0, R1', R2', R3', R4'. To predict whether the reducing monosaccharide will be covalently linked to the central residue R0(Ser or Thr), a new method based on feature selection has been proposed in our work. 277 nonapeptides from reference [Chou KC. A sequence-coupled vector-projection model for predicting the specificity of GalNAc-transferase. Protein Sci 1995;4:1365-83] are chosen for training set. Each nonapeptide is represented by hundreds of amino acid properties collected by Amino Acid Index database (http://www.genome.jp/aaindex) and transformed into a numeric vector with 4554 features. The Maximum Relevance Minimum Redundancy (mRMR) method combining with Incremental Feature Selection (IFS) and Feature Forward Selection (FFS) are then applied for feature selection. Nearest Neighbor Algorithm (NNA) is used to build prediction models. The optimal model contains 54 features and its correct rate tested by Jackknife cross-validation test reaches 91.34%. Final feature analysis indicates that amino acid residues at position R3' play the most important role in the recognition of GalNAc-transferase specificity, which were confirmed by the experiments [Elhammer AP, Poorman RA, Brown E, Maggiora LL, Hoogerheide JG, Kezdy FJ. The specificity of UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase as inferred from a database of in vivo substrates and from the in vitro glycosylation of proteins and peptides. J Biol Chem 1993;268:10029-38; O'Connell BC, Hagen FK, Tabak LA. The influence of flanking sequence on the O-glycosylation of threonine in vitro. J Biol Chem 1992;267:25010-8; Yoshida A, Suzuki M, Ikenaga H, Takeuchi M. Discovery of the shortest sequence motif for high level mucin-type O-glycosylation. J Biol Chem 1997;272:16884-8]. Our method can be used as a tool for predicting O-glycosylation sites and for investigating the GalNAc-transferase specificity, which is useful for designing competitive inhibitors of GalNAc-transferase. The predicting software is available upon the request.     

Feature selection for splice site prediction: A new method using EDA-based feature ranking

Background  

The identification of relevant biological features in large and complex datasets is an important step towards gaining insight in the processes underlying the data. Other advantages of feature selection include the ability of the classification system to attain good or even better solutions using a restricted subset of features, and a faster classification. Thus, robust methods for fast feature selection are of key importance in extracting knowledge from complex biological data.     

Tumor classification and marker gene prediction by feature selection and fuzzy c-means clustering using microarray data

Background  

Using DNA microarrays, we have developed two novel models for tumor classification and target gene prediction. First, gene expression profiles are summarized by optimally selected Self-Organizing Maps (SOMs), followed by tumor sample classification by Fuzzy C-means clustering. Then, the prediction of marker genes is accomplished by either manual feature selection (visualizing the weighted/mean SOM component plane) or automatic feature selection (by pair-wise Fisher's linear discriminant).     

FrankSum: new feature selection method for protein function prediction

In the study of in silico functional genomics, improving the performance of protein function prediction is the ultimate goal for identifying proteins associated with defined cellular functions. The classical prediction approach is to employ pairwise sequence alignments. However this method often faces difficulties when no statistically significant homologous sequences are identified. An alternative way is to predict protein function from sequence-derived features using machine learning. In this case the choice of possible features which can be derived from the sequence is of vital importance to ensure adequate discrimination to predict function. In this paper we have successfully selected biologically significant features for protein function prediction. This was performed using a new feature selection method (FrankSum) that avoids data distribution assumptions, uses a data independent measurement (p-value) within the feature, identifies redundancy between features and uses an appropriate ranking criterion for feature selection. We have shown that classifiers generated from features selected by FrankSum outperforms classifiers generated from full feature sets, randomly selected features and features selected from the Wrapper method. We have also shown the features are concordant across all species and top ranking features are biologically informative. We conclude that feature selection is vital for successful protein function prediction and FrankSum is one of the feature selection methods that can be applied successfully to such a domain.     

Sequence-based classification using discriminatory motif feature selection

Most existing methods for sequence-based classification use exhaustive feature generation, employing, for example, all k-mer patterns. The motivation behind such (enumerative) approaches is to minimize the potential for overlooking important features. However, there are shortcomings to this strategy. First, practical constraints limit the scope of exhaustive feature generation to patterns of length ≤ k, such that potentially important, longer (> k) predictors are not considered. Second, features so generated exhibit strong dependencies, which can complicate understanding of derived classification rules. Third, and most importantly, numerous irrelevant features are created. These concerns can compromise prediction and interpretation. While remedies have been proposed, they tend to be problem-specific and not broadly applicable. Here, we develop a generally applicable methodology, and an attendant software pipeline, that is predicated on discriminatory motif finding. In addition to the traditional training and validation partitions, our framework entails a third level of data partitioning, a discovery partition. A discriminatory motif finder is used on sequences and associated class labels in the discovery partition to yield a (small) set of features. These features are then used as inputs to a classifier in the training partition. Finally, performance assessment occurs on the validation partition. Important attributes of our approach are its modularity (any discriminatory motif finder and any classifier can be deployed) and its universality (all data, including sequences that are unaligned and/or of unequal length, can be accommodated). We illustrate our approach on two nucleosome occupancy datasets and a protein solubility dataset, previously analyzed using enumerative feature generation. Our method achieves excellent performance results, with and without optimization of classifier tuning parameters. A Python pipeline implementing the approach is available at http://www.epibiostat.ucsf.edu/biostat/sen/dmfs/.     

Drug permeability prediction using PMF method

Drug permeability determines the oral availability of drugs via cellular membranes. Poor permeability makes a drug unsuitable for further development. The permeability may be estimated as the free energy change that the drug should overcome through crossing membrane. In this paper the drug permeability was simulated using molecular dynamics method and the potential energy profile was calculated with potential of mean force (PMF) method. The membrane was simulated using DPPC bilayer and three drugs with different permeability were tested. PMF studies on these three drugs show that doxorubicin (low permeability) should pass higher free energy barrier from water to DPPC bilayer center while ibuprofen (high permeability) has a lower energy barrier. Our calculation indicates that the simulation model we built is suitable to predict drug permeability.     

Ancestral informative marker selection and population structure visualization using sparse Laplacian eigenfunctions

Identification of a small panel of population structure informative markers can reduce genotyping cost and is useful in various applications, such as ancestry inference in association mapping, forensics and evolutionary theory in population genetics. Traditional methods to ascertain ancestral informative markers usually require the prior knowledge of individual ancestry and have difficulty for admixed populations. Recently Principal Components Analysis (PCA) has been employed with success to select SNPs which are highly correlated with top significant principal components (PCs) without use of individual ancestral information. The approach is also applicable to admixed populations. Here we propose a novel approach based on our recent result on summarizing population structure by graph laplacian eigenfunctions, which differs from PCA in that it is geometric and robust to outliers. Our approach also takes advantage of the priori sparseness of informative markers in the genome. Through simulation of a ring population and the real global population sample HGDP of 650K SNPs genotyped in 940 unrelated individuals, we validate the proposed algorithm at selecting most informative markers, a small fraction of which can recover the similar underlying population structure efficiently. Employing a standard Support Vector Machine (SVM) to predict individuals' continental memberships on HGDP dataset of seven continents, we demonstrate that the selected SNPs by our method are more informative but less redundant than those selected by PCA. Our algorithm is a promising tool in genome-wide association studies and population genetics, facilitating the selection of structure informative markers, efficient detection of population substructure and ancestral inference.     

Prediction of thermophilic proteins using feature selection technique

The thermostability of proteins is particularly relevant for enzyme engineering. Developing a computational method to identify mesophilic proteins would be helpful for protein engineering and design. In this work, we developed support vector machine based method to predict thermophilic proteins using the information of amino acid distribution and selected amino acid pairs. A reliable benchmark dataset including 915 thermophilic proteins and 793 non-thermophilic proteins was constructed for training and testing the proposed models. Results showed that 93.8% thermophilic proteins and 92.7% non-thermophilic proteins could be correctly predicted by using jackknife cross-validation. High predictive successful rate exhibits that this model can be applied for designing stable proteins.     

A novel sequence-based method for phosphorylation site prediction with feature selection and analysis

Phosphorylation is one of the most important post-translational modifications, and the identification of protein phosphorylation sites is particularly important for studying disease diagnosis. However, experimental detection of phosphorylation sites is labor intensive. It would be beneficial if computational methods are available to provide an extra reference for the phosphorylation sites. Here we developed a novel sequence-based method for serine, threonine, and tyrosine phosphorylation site prediction. Nearest Neighbor algorithm was employed as the prediction engine. The peptides around the phosphorylation sites with a fixed length of thirteen amino acid residues were extracted via a sliding window along the protein chains concerned. Each of such peptides was coded into a vector with 6,072 features, derived from Amino Acid Index (AAIndex) database, for the classification/detection. Incremental Feature Selection, a feature selection algorithm based on the Maximum Relevancy Minimum Redundancy (mRMR) method was used to select a compact feature set for a further improvement of the classification performance. Three predictors were established for identifying the three types of phosphorylation sites, achieving the overall accuracies of 66.64%, 66.11%% and 66.69%, respectively. These rates were obtained by rigorous jackknife cross-validation tests.     

Iterative class discovery and feature selection using Minimal Spanning Trees

Background  

Clustering is one of the most commonly used methods for discovering hidden structure in microarray gene expression data. Most current methods for clustering samples are based on distance metrics utilizing all genes. This has the effect of obscuring clustering in samples that may be evident only when looking at a subset of genes, because noise from irrelevant genes dominates the signal from the relevant genes in the distance calculation.     

Fast feature selection using a simple estimation of distribution algorithm: a case study on splice site prediction

MOTIVATION: Feature subset selection is an important preprocessing step for classification. In biology, where structures or processes are described by a large number of features, the elimination of irrelevant and redundant information in a reasonable amount of time has a number of advantages. It enables the classification system to achieve good or even better solutions with a restricted subset of features, allows for a faster classification, and it helps the human expert focus on a relevant subset of features, hence providing useful biological knowledge. RESULTS: We present a heuristic method based on Estimation of Distribution Algorithms to select relevant subsets of features for splice site prediction in Arabidopsis thaliana. We show that this method performs a fast detection of relevant feature subsets using the technique of constrained feature subsets. Compared to the traditional greedy methods the gain in speed can be up to one order of magnitude, with results being comparable or even better than the greedy methods. This makes it a very practical solution for classification tasks that can be solved using a relatively small amount of discriminative features (or feature dependencies), but where the initial set of potential discriminative features is rather large.     

Calpain cleavage prediction using multiple kernel learning

Calpain, an intracellular -dependent cysteine protease, is known to play a role in a wide range of metabolic pathways through limited proteolysis of its substrates. However, only a limited number of these substrates are currently known, with the exact mechanism of substrate recognition and cleavage by calpain still largely unknown. While previous research has successfully applied standard machine-learning algorithms to accurately predict substrate cleavage by other similar types of proteases, their approach does not extend well to calpain, possibly due to its particular mode of proteolytic action and limited amount of experimental data. Through the use of Multiple Kernel Learning, a recent extension to the classic Support Vector Machine framework, we were able to train complex models based on rich, heterogeneous feature sets, leading to significantly improved prediction quality (6% over highest AUC score produced by state-of-the-art methods). In addition to producing a stronger machine-learning model for the prediction of calpain cleavage, we were able to highlight the importance and role of each feature of substrate sequences in defining specificity: primary sequence, secondary structure and solvent accessibility. Most notably, we showed there existed significant specificity differences across calpain sub-types, despite previous assumption to the contrary. Prediction accuracy was further successfully validated using, as an unbiased test set, mutated sequences of calpastatin (endogenous inhibitor of calpain) modified to no longer block calpain''s proteolytic action. An online implementation of our prediction tool is available at http://calpain.org.     

Automated skin biopsy histopathological image annotation using multi-instance representation and learning

With digitisation and the development of computer-aided diagnosis, histopathological image analysis has attracted considerable interest in recent years. In this article, we address the problem of the automated annotation of skin biopsy images, a special type of histopathological image analysis. In contrast to previous well-studied methods in histopathology, we propose a novel annotation method based on a multi-instance learning framework. The proposed framework first represents each skin biopsy image as a multi-instance sample using a graph cutting method, decomposing the image to a set of visually disjoint regions. Then, we construct two classification models using multi-instance learning algorithms, among which one provides determinate results and the other calculates a posterior probability. We evaluate the proposed annotation framework using a real dataset containing 6691 skin biopsy images, with 15 properties as target annotation terms. The results indicate that the proposed method is effective and medically acceptable.