DeepAdd: Protein function prediction from k-mer embedding and additional features

With the application of new high throughput sequencing technology, a large number of protein sequences is becoming available. Determination of the functional characteristics of these proteins by experiments is an expensive endeavor that requires a lot of time. Furthermore, at the organismal level, such kind of experimental functional analyses can be conducted only for a very few selected model organisms. Computational function prediction methods can be used to fill this gap. The functions of proteins are classified by Gene Ontology (GO), which contains more than 40,000 classifications in three domains, Molecular Function (MF), Biological Process (BP), and Cellular Component (CC). Additionally, since proteins have many functions, function prediction represents a multi-label and multi-class problem. We developed a new method to predict protein function from sequence. To this end, natural language model was used to generate word embedding of sequence and learn features from it by deep learning, and additional features to locate every protein. Our method uses the dependencies between GO classes as background information to construct a deep learning model. We evaluate our method using the standards established by the Computational Assessment of Function Annotation (CAFA) and have noticeable improvement over several algorithms, such as FFPred, DeepGO, GoFDR and other methods compared on the CAFA3 datasets.     

Ensembling machine learning models to boost molecular affinity prediction

This study unites six popular machine learning approaches to enhance the prediction of a molecular binding affinity between receptors (large protein molecules) and ligands (small organic molecules). Here we examine a scheme where affinity of ligands is predicted against a single receptor – human thrombin, thus, the models consider ligand features only. However, the suggested approach can be repurposed for other receptors. The methods include Support Vector Machine, Random Forest, CatBoost, feed-forward neural network, graph neural network, and Bidirectional Encoder Representations from Transformers. The first five methods use input features based on physico-chemical properties of molecules, while the last one is based on textual molecular representations. All approaches do not rely on atomic spatial coordinates, avoiding a potential bias from known structures, and are capable of generalizing for compounds with unknown conformations. Within each of the methods, we have trained two models that solve classification and regression tasks. Then, all models are grouped into a pipeline of two subsequent ensembles. The first ensemble aggregates six classification models which vote whether a ligand binds to a receptor or not. If a ligand is classified as active (i.e., binds), the second ensemble predicts its binding affinity in terms of the inhibition constant K_i.     

LMI-DForest: A deep forest model towards the prediction of lncRNA-miRNA interactions

The interactions between miRNAs and long non-coding RNAs (lncRNAs) are subject to intensive recent studies due to its critical role in gene regulations. Computational prediction of lncRNA-miRNA interactions has become a popular alternative strategy to the experimental methods for identification of underlying interactions. It is desirable to develop the machine learning-based models for prediction of lncRNA-miRNA based on the experimentally validated interactions between lncRNAs and miRNAs. The accuracy and robustness of existing models based on machine learning techniques are subject to further improvement.Considering that the attributes of lncRNA and miRNA contribute key importance in the interaction between these two RNAs, a deep learning model, named LMI-DForest, is proposed here by combining the deep forest and autoencoder strategies. Systematic comparison on the experiment validated datasets for lncRNA-miRNA interaction datasets demonstrates that the proposed method consistently shows superior performance over the other machine learning models in the lncRNA-miRNA interaction prediction.     

CE-PLoc: an ensemble classifier for predicting protein subcellular locations by fusing different modes of pseudo amino acid composition

Precise information about protein locations in a cell facilitates in the understanding of the function of a protein and its interaction in the cellular environment. This information further helps in the study of the specific metabolic pathways and other biological processes. We propose an ensemble approach called "CE-PLoc" for predicting subcellular locations based on fusion of individual classifiers. The proposed approach utilizes features obtained from both dipeptide composition (DC) and amphiphilic pseudo amino acid composition (PseAAC) based feature extraction strategies. Different feature spaces are obtained by varying the dimensionality using PseAAC for a selected base learner. The performance of the individual learning mechanisms such as support vector machine, nearest neighbor, probabilistic neural network, covariant discriminant, which are trained using PseAAC based features is first analyzed. Classifiers are developed using same learning mechanism but trained on PseAAC based feature spaces of varying dimensions. These classifiers are combined through voting strategy and an improvement in prediction performance is achieved. Prediction performance is further enhanced by developing CE-PLoc through the combination of different learning mechanisms trained on both DC based feature space and PseAAC based feature spaces of varying dimensions. The predictive performance of proposed CE-PLoc is evaluated for two benchmark datasets of protein subcellular locations using accuracy, MCC, and Q-statistics. Using the jackknife test, prediction accuracies of 81.47 and 83.99% are obtained for 12 and 14 subcellular locations datasets, respectively. In case of independent dataset test, prediction accuracies are 87.04 and 87.33% for 12 and 14 class datasets, respectively.     

Machine learning algorithms for predicting protein folding rates and stability of mutant proteins: comparison with statistical methods

Machine learning algorithms have wide range of applications in bioinformatics and computational biology such as prediction of protein secondary structures, solvent accessibility, binding site residues in protein complexes, protein folding rates, stability of mutant proteins, and discrimination of proteins based on their structure and function. In this work, we focus on two aspects of predictions: (i) protein folding rates and (ii) stability of proteins upon mutations. We briefly introduce the concepts of protein folding rates and stability along with available databases, features for prediction methods and measures for prediction performance. Subsequently, the development of structure based parameters and their relationship with protein folding rates will be outlined. The structure based parameters are helpful to understand the physical basis for protein folding and stability. Further, basic principles of major machine learning techniques will be mentioned and their applications for predicting protein folding rates and stability of mutant proteins will be illustrated. The machine learning techniques could achieve the highest accuracy of predicting protein folding rates and stability. In essence, statistical methods and machine learning algorithms are complimenting each other for understanding and predicting protein folding rates and the stability of protein mutants. The available online resources on protein folding rates and stability will be listed.     

Application of deep eutectic solvents in the extraction and separation of target compounds from various samples

Deep eutectic solvents, as a new type of eco‐friendly solvent, have attracted increasing attention in chemistry for the extraction and separation of target compounds from various samples. To summarize the application of deep eutectic solvents, this review highlights some of the unique properties of deep eutectic solvents and deep‐eutectic‐solvent‐based materials, as well as their applications in extraction and separation. In this paper, the available data and references in this field are reviewed to summarize the application developments of deep eutectic solvents. Based on the development of deep eutectic solvents, the exploitation of new deep eutectic solvents and deep‐eutectic‐solvent‐based materials are expected to diversify into extraction and separation.     

A representation transfer learning approach for enhanced prediction of growth hormone binding proteins

Growth hormone binding proteins (GHBPs) are soluble proteins that play an important role in the modulation of signaling pathways pertaining to growth hormones. GHBPs are selective and bind non-covalently with growth hormones, but their functions are still not fully understood. Identification and characterization of GHBPs are the preliminary steps for understanding their roles in various cellular processes. As wet lab based experimental methods involve high cost and labor, computational methods can facilitate in narrowing down the search space of putative GHBPs. Performance of machine learning algorithms largely depends on the quality of features that it feeds on. Informative and non-redundant features generally result in enhanced performance and for this purpose feature selection algorithms are commonly used. In the present work, a novel representation transfer learning approach is presented for prediction of GHBPs. For their accurate prediction, deep autoencoder based features were extracted and subsequently SMO-PolyK classifier is trained. The prediction model is evaluated by both leave one out cross validation (LOOCV) and hold out independent testing set. On LOOCV, the prediction model achieved 89.8%% accuracy, with 89.4% sensitivity and 90.2% specificity and accuracy of 93.5%, sensitivity of 90.2% and specificity of 96.8% is attained on the hold out testing set. Further a comparison was made between the full set of sequence-based features, top performing sequence features extracted using feature selection algorithm, deep autoencoder based features and generalized low rank model based features on the prediction accuracy. Principal component analysis of the representative features along with t-sne visualization demonstrated the effectiveness of deep features in prediction of GHBPs. The present method is robust and accurate and may complement other wet lab based methods for identification of novel GHBPs.     

Protein function prediction using neighbor relativity in protein–protein interaction network

There is a large gap between the number of discovered proteins and the number of functionally annotated ones. Due to the high cost of determining protein function by wet-lab research, function prediction has become a major task for computational biology and bioinformatics. Some researches utilize the proteins interaction information to predict function for un-annotated proteins. In this paper, we propose a novel approach called “Neighbor Relativity Coefficient” (NRC) based on interaction network topology which estimates the functional similarity between two proteins. NRC is calculated for each pair of proteins based on their graph-based features including distance, common neighbors and the number of paths between them. In order to ascribe function to an un-annotated protein, NRC estimates a weight for each neighbor to transfer its annotation to the unknown protein. Finally, the unknown protein will be annotated by the top score transferred functions. We also investigate the effect of using different coefficients for various types of functions. The proposed method has been evaluated on Saccharomyces cerevisiae and Homo sapiens interaction networks. The performance analysis demonstrates that NRC yields better results in comparison with previous protein function prediction approaches that utilize interaction network.     

Genome Mining and Molecular Networking-Based Metabolomics of the Marine Facultative Aspergillus sp. MEXU 27854

The marine-facultative Aspergillus sp. MEXU 27854, isolated from the Caleta Bay in Acapulco, Guerrero, Mexico, has provided an interesting diversity of secondary metabolites, including a series of rare dioxomorpholines, peptides, and butyrolactones. Here, we report on the genomic data, which consists of 11 contigs (N50~3.95 Mb) with a ~30.75 Mb total length of assembly. Genome annotation resulted in the prediction of 10,822 putative genes. Functional annotation was accomplished by BLAST searching protein sequences with different public databases. Of the predicted genes, 75% were assigned gene ontology terms. From the 67 BGCs identified, ~60% belong to the NRPS and NRPS-like classes. Putative BGCs for the dioxomorpholines and other metabolites were predicted by extensive genome mining. In addition, metabolomic molecular networking analysis allowed the annotation of all isolated compounds and revealed the biosynthetic potential of this fungus. This work represents the first report of whole-genome sequencing and annotation from a marine-facultative fungal strain isolated from Mexico.     

Application of Intelligent Techniques for Classification of Bacteria Using Protein Sequence-Derived Features

Standard molecular experimental methodologies and mathematical procedures often fail to answer many phylogeny and classification related issues. Modern artificial intelligent-based techniques, such as radial basis function, genetic algorithm, artificial neural network, and support vector machines are of ample potential in this regard. Reliance on a large number of essential parameters will aid in enhanced robustness, reliability, and better accuracy as opposed to single molecular parameter. This study was conducted with dataset of computed protein physicochemical properties belonging to 20 different bacterial genera. A total of 57 sequential and structural parameters derived from protein sequences were considered for the initial classification. Feature selection based techniques were employed to find out the most important features influencing the dataset. Various amino acids, hydrophobicity, relative sulfur percentage, and codon number were selected as important parameters during the study. Comparative analyses were performed applying RapidMiner data mining platform. Support vector machine proved to be the best method with maximum accuracy of more than 91 %.     

FunCat functional inference with belief propagation and feature integration

Pairwise comparison of sequence data is intensively used for automated functional protein annotation, while graphical models emerge as promising candidates for an integration of various heterogeneous features. We designed a model, termed hRMN that integrates different genomic features and implemented a variant of belief propagation for functional annotation transfer. hRMN allows the assignment of multiple functional categories while avoiding common problems in annotation transfer from heterogeneous datasets, such as an independency of the investigated datasets. We benchmarked this system with large-scale annotation transfer (based on the MIPS FunCat ontology) to proteins of the prokaryotes Bacillus subtilis, Helicobacter pylori, Listeria monocytogenes, and Listeria innocua. hRMN consistently outperformed two competitors in annotation of four bacterial genomes. The developed code is available for download at http://mips.gsf.de/proj/bfab/hRMN.html.     

Self-supervised clustering of mass spectrometry imaging data using contrastive learning

Mass spectrometry imaging (MSI) is widely used for the label-free molecular mapping of biological samples. The identification of co-localized molecules in MSI data is crucial to the understanding of biochemical pathways. One of key challenges in molecular colocalization is that complex MSI data are too large for manual annotation but too small for training deep neural networks. Herein, we introduce a self-supervised clustering approach based on contrastive learning, which shows an excellent performance in clustering of MSI data. We train a deep convolutional neural network (CNN) using MSI data from a single experiment without manual annotations to effectively learn high-level spatial features from ion images and classify them based on molecular colocalizations. We demonstrate that contrastive learning generates ion image representations that form well-resolved clusters. Subsequent self-labeling is used to fine-tune both the CNN encoder and linear classifier based on confidently classified ion images. This new approach enables autonomous and high-throughput identification of co-localized species in MSI data, which will dramatically expand the application of spatial lipidomics, metabolomics, and proteomics in biological research.

Contrastive learning is used to train a deep convolutional neural network to identify high-level features in mass spectrometry imaging data. These features enable self-supervised clustering of ion images without manual annotation.     

TargetATPsite: A template‐free method for ATP‐binding sites prediction with residue evolution image sparse representation and classifier ensemble

Understanding the interactions between proteins and ligands is critical for protein function annotations and drug discovery. We report a new sequence‐based template‐free predictor (TargetATPsite) to identify the Adenosine‐5′‐triphosphate (ATP) binding sites with machine‐learning approaches. Two steps are implemented in TargetATPsite: binding residues and pockets predictions, respectively. To predict the binding residues, a novel image sparse representation technique is proposed to encode residue evolution information treated as the input features. An ensemble classifier constructed based on support vector machines (SVM) from multiple random under‐samplings is used as the prediction model, which is effective for dealing with imbalance phenomenon between the positive and negative training samples. Compared with the existing ATP‐specific sequence‐based predictors, TargetATPsite is featured by the second step of possessing the capability of further identifying the binding pockets from the predicted binding residues through a spatial clustering algorithm. Experimental results on three benchmark datasets demonstrate the efficacy of TargetATPsite. © 2013 Wiley Periodicals, Inc.     

Effective single-cell clustering through ensemble feature selection and similarity measurements

Single-cell RNA sequencing technologies have revolutionized biomedical research by providing an effective means to profile gene expressions in individual cells. One of the first fundamental steps to perform the in-depth analysis of single-cell sequencing data is cell type classification and identification. Computational methods such as clustering algorithms have been utilized and gaining in popularity because they can save considerable resources and time for experimental validations. Although selecting the optimal features (i.e., genes) is an essential process to obtain accurate and reliable single-cell clustering results, the computational complexity and dropout events that can introduce zero-inflated noise make this process very challenging. In this paper, we propose an effective single-cell clustering algorithm based on the ensemble feature selection and similarity measurements. We initially identify the set of potential features, then measure the cell-to-cell similarity based on the subset of the potentials through multiple feature sampling approaches. We construct the ensemble network based on cell-to-cell similarity. Finally, we apply a network-based clustering algorithm to obtain single-cell clusters. We evaluate the performance of our proposed algorithm through multiple assessments in real-world single-cell RNA sequencing datasets with known cell types. The results show that our proposed algorithm can identify accurate and consistent single-cell clustering. Moreover, the proposed algorithm takes relative expression as input, so it can easily be adopted by existing analysis pipelines. The source code has been made publicly available at https://github.com/jeonglab/scCLUE.     

In silico Protein-Protein Interaction prediction with sequence alignment and classifier stacking

Protein-Protein Interaction (PPI) prediction is a well known problem in Bioinformatics, for which a large number of techniques have been proposed in the past. However, prediction results have not been sufficiently satisfactory for guiding biologists in web-lab experiments. One reason is that not all useful information, such as pairwise protein interaction information based on sequence alignment, has been integrated together in PPI prediction. Alignment is a basic concept to measure sequence similarity in Proteomics that has been used in a number of applications ranging from protein recognition to protein subcellular localization. In this article, we propose a novel integrated approach to predicting PPI based on sequence alignment by jointly using a k-Nearest Neighbor classifier (SA-kNN) and a Support Vector Machine (SVM). SVM is a machine learning technique used in a wide range of Bioinformatics applications, thanks to the ability to alleviate the overfitting problems. We demonstrate that in our approach the two methods, SA-kNN and SVM, are complementary, which are combined in an ensemble to overcome their respective limitations. While the SVM is trained on Amino Acid (AA) compositions and protein signatures mined from literature, the SA-kNN makes use of the similarity of two protein pairs through alignment. Experimentally, our technique leads to a significant gain in accuracy, precision and sensitivity measures at ~5%, 16% and 10% respectively.     

Multi-instance multi-label distance metric learning for genome-wide protein function prediction

Multi-instance multi-label (MIML) learning has been proven to be effective for the genome-wide protein function prediction problems where each training example is associated with not only multiple instances but also multiple class labels. To find an appropriate MIML learning method for genome-wide protein function prediction, many studies in the literature attempted to optimize objective functions in which dissimilarity between instances is measured using the Euclidean distance. But in many real applications, Euclidean distance may be unable to capture the intrinsic similarity/dissimilarity in feature space and label space. Unlike other previous approaches, in this paper, we propose to learn a multi-instance multi-label distance metric learning framework (MIMLDML) for genome-wide protein function prediction. Specifically, we learn a Mahalanobis distance to preserve and utilize the intrinsic geometric information of both feature space and label space for MIML learning. In addition, we try to deal with the sparsely labeled data by giving weight to the labeled data. Extensive experiments on seven real-world organisms covering the biological three-domain system (i.e., archaea, bacteria, and eukaryote; Woese et al., 1990) show that the MIMLDML algorithm is superior to most state-of-the-art MIML learning algorithms.