Protein remote homology detection based on bidirectional long short-term memory

Classification of DNA Sequence Based on a Non-gradient Algorithm: Pseudoinverse Learners

This chapter proposes a prototype-based classification approach for analyzing DNA barcodes that uses a spectral representation of DNA sequences and a non-gradient neural network. Biological sequences can be viewed as data components with higher non-fixed dimensions, which correspond to the length of the sequences. Through computational procedures such as one-hot encoding, numerical encoding plays an important role in DNA sequence evaluation (OHE). However, the OHE method has some disadvantages: (1) It does not add any details that could result in an additional predictive variable, and (2) if the variable has many classes, OHE significantly expands the feature space. To address these shortcomings, this chapter proposes a computationally efficient framework for classifying DNA sequences of living organisms in the image domain. A multilayer perceptron trained by a pseudoinverse learning autoencoder (PILAE) algorithm is used in the proposed strategy. The learning control parameters and the number of hidden layers do not have to be specified during the PILAE training process. As a result, the PILAE classifier outperforms other deep neural network (DNN) strategies such as the VGG-16 and Xception models.

Deep Learning for Genomics: From Early Neural Nets to Modern Large Language Models

The data explosion driven by advancements in genomic research, such as high-throughput sequencing techniques, is constantly challenging conventional methods used in genomics. In parallel with the urgent demand for robust algorithms, deep learning has succeeded in various fields such as vision, speech, and text processing. Yet genomics entails unique challenges to deep learning, since we expect a superhuman intelligence that explores beyond our knowledge to interpret the genome from deep learning. A powerful deep learning model should rely on the insightful utilization of task-specific knowledge. In this paper, we briefly discuss the strengths of different deep learning models from a genomic perspective so as to fit each particular task with proper deep learning-based architecture, and we remark on practical considerations of developing deep learning architectures for genomics. We also provide a concise review of deep learning applications in various aspects of genomic research and point out current challenges and potential research directions for future genomics applications. We believe the collaborative use of ever-growing diverse data and the fast iteration of deep learning models will continue to contribute to the future of genomics.

pLM-BLAST: distant homology detection based on direct comparison of sequence representations from protein language models

MOTIVATION

The detection of homology through sequence comparison is a typical first step in the study of protein function and evolution. In this work, we explore the applicability of protein language models to this task.

RESULTS

We introduce pLM-BLAST, a tool inspired by BLAST, that detects distant homology by comparing single-sequence representations (embeddings) derived from a protein language model, ProtT5. Our benchmarks reveal that pLM-BLAST maintains a level of accuracy on par with HHsearch for both highly similar sequences (with >50% identity) and markedly divergent sequences (with <30% identity), while being significantly faster. Additionally, pLM-BLAST stands out among other embedding-based tools due to its ability to compute local alignments. We show that these local alignments, produced by pLM-BLAST, often connect highly divergent proteins, thereby highlighting its potential to uncover previously undiscovered homologous relationships and improve protein annotation.

AVAILABILITY AND IMPLEMENTATION

pLM-BLAST is accessible via the MPI Bioinformatics Toolkit as a web server for searching precomputed databases (https://toolkit.tuebingen.mpg.de/tools/plmblast). It is also available as a standalone tool for building custom databases and performing batch searches (https://github.com/labstructbioinf/pLM-BLAST).

Principles and challenges of modeling temporal and spatial omics data

Studies with temporal or spatial resolution are crucial to understand the molecular dynamics and spatial dependencies underlying a biological process or system. With advances in high-throughput omic technologies, time- and space-resolved molecular measurements at scale are increasingly accessible, providing new opportunities to study the role of timing or structure in a wide range of biological questions. At the same time, analyses of the data being generated in the context of spatiotemporal studies entail new challenges that need to be considered, including the need to account for temporal and spatial dependencies and compare them across different scales, biological samples or conditions. In this Review, we provide an overview of common principles and challenges in the analysis of temporal and spatial omics data. We discuss statistical concepts to model temporal and spatial dependencies and highlight opportunities for adapting existing analysis methods to data with temporal and spatial dimensions.

PreHom-PCLM: protein remote homology detection by combing motifs and protein cubic language model

Protein remote homology detection is essential for structure prediction, function prediction, disease mechanism understanding, etc. The remote homology relationship depends on multiple protein properties, such as structural information and local sequence patterns. Previous studies have shown the challenges for predicting remote homology relationship by protein features at sequence level (e.g. position-specific score matrix). Protein motifs have been used in structure and function analysis due to their unique sequence patterns and implied structural information. Therefore, designing a usable architecture to fuse multiple protein properties based on motifs is urgently needed to improve protein remote homology detection performance. To make full use of the characteristics of motifs, we employed the language model called the protein cubic language model (PCLM). It combines multiple properties by constructing a motif-based neural network. Based on the PCLM, we proposed a predictor called PreHom-PCLM by extracting and fusing multiple motif features for protein remote homology detection. PreHom-PCLM outperforms the other state-of-the-art methods on the test set and independent test set. Experimental results further prove the effectiveness of multiple features fused by PreHom-PCLM for remote homology detection. Furthermore, the protein features derived from the PreHom-PCLM show strong discriminative power for proteins from different structural classes in the high-dimensional space. Availability and Implementation: http://bliulab.net/PreHom-PCLM.

Machine Learning Methods for Small Data Challenges in Molecular Science

Small data are often used in scientific and engineering research due to the presence of various constraints, such as time, cost, ethics, privacy, security, and technical limitations in data acquisition. However, big data have been the focus for the past decade, small data and their challenges have received little attention, even though they are technically more severe in machine learning (ML) and deep learning (DL) studies. Overall, the small data challenge is often compounded by issues, such as data diversity, imputation, noise, imbalance, and high-dimensionality. Fortunately, the current big data era is characterized by technological breakthroughs in ML, DL, and artificial intelligence (AI), which enable data-driven scientific discovery, and many advanced ML and DL technologies developed for big data have inadvertently provided solutions for small data problems. As a result, significant progress has been made in ML and DL for small data challenges in the past decade. In this review, we summarize and analyze several emerging potential solutions to small data challenges in molecular science, including chemical and biological sciences. We review both basic machine learning algorithms, such as linear regression, logistic regression (LR), k-nearest neighbor (KNN), support vector machine (SVM), kernel learning (KL), random forest (RF), and gradient boosting trees (GBT), and more advanced techniques, including artificial neural network (ANN), convolutional neural network (CNN), U-Net, graph neural network (GNN), Generative Adversarial Network (GAN), long short-term memory (LSTM), autoencoder, transformer, transfer learning, active learning, graph-based semi-supervised learning, combining deep learning with traditional machine learning, and physical model-based data augmentation. We also briefly discuss the latest advances in these methods. Finally, we conclude the survey with a discussion of promising trends in small data challenges in molecular science.

Deep Neural Network Framework Based on Word Embedding for Protein Glutarylation Sites Prediction

In recent years, much research has found that dysregulation of glutarylation is associated with many human diseases, such as diabetes, cancer, and glutaric aciduria type I. Therefore, glutarylation identification and characterization are essential tasks for determining modification-specific proteomics. This study aims to propose a novel deep neural network framework based on word embedding techniques for glutarylation sites prediction. Multiple deep neural network models are implemented to evaluate the performance of glutarylation sites prediction. Furthermore, an extensive experimental comparison of word embedding techniques is conducted to utilize the most efficient method for improving protein sequence data representation. The results suggest that the proposed deep neural networks not only improve protein sequence representation but also work effectively in glutarylation sites prediction by obtaining a higher accuracy and confidence rate compared to the previous work. Moreover, embedding techniques were proven to be more productive than the pre-trained word embedding techniques for glutarylation sequence representation. Our proposed method has significantly outperformed all traditional performance metrics compared to the advanced integrated vector support, with accuracy, specificity, sensitivity, and correlation coefficient of 0.79, 0.89, 0.59, and 0.51, respectively. It shows the potential to detect new glutarylation sites and uncover the relationships between glutarylation and well-known lysine modification.

Evolutionary analyses of genes in Echinodermata offer insights towards the origin of metazoan phyla

Despite recent studies discussing the evolutionary impacts of gene duplications and losses among metazoans, the genomic basis for the evolution of phyla remains enigmatic. Here, we employ phylogenomic approaches to search for orthologous genes without known functions among echinoderms, and subsequently use them to guide the identification of their homologs across other metazoans. Our final set of 14 genes was obtained via a suite of homology prediction tools, gene expression data, gene ontology, and generating the Strongylocentrotus purpuratus phylome. The gene set was subjected to selection pressure analyses, which indicated that they are highly conserved and under negative selection. Their presence across broad taxonomic depths suggests that genes required to form a phylum are ancestral to that phylum. Therefore, rather than de novo gene genesis, we posit that evolutionary forces such as selection on existing genomic elements over large timescales may drive divergence and contribute to the emergence of phyla.

Computational analysis and prediction of PE_PGRS proteins using machine learning

•

PEPPER is the first machine learning-based predictor for PE_PGRS proteins.

•

PEPPER is based on lightGBM and various sequence and physicochemical features.

•

PEPPER can identify PE_PGRS proteins rapidly and accurately.

•

The webserver of PEPPER and stand-alone tool are publicly available at http://web.unimelb-bioinfortools.cloud.edu.au/PEPPER/.

Mycobacterium tuberculosis genome comprises approximately 10% of two families of poorly characterised genes due to their high GC content and highly repetitive nature. The largest sub-group, the proline-glutamic acid polymorphic guanine-cytosine-rich sequence (PE_PGRS) family, is thought to be involved in host response and disease pathogenicity. Due to their high genetic variability and complexity of analysis, they are typically disregarded for further research in genomic studies. There are currently limited online resources and homology computational tools that can identify and analyse PE_PGRS proteins. In addition, they are computational-intensive and time-consuming, and lack sensitivity. Therefore, computational methods that can rapidly and accurately identify PE_PGRS proteins are valuable to facilitate the functional elucidation of the PE_PGRS family proteins. In this study, we developed the first machine learning-based bioinformatics approach, termed PEPPER, to allow users to identify PE_PGRS proteins rapidly and accurately. PEPPER was built upon a comprehensive evaluation of 13 popular machine learning algorithms with various sequence and physicochemical features. Empirical studies demonstrated that PEPPER achieved significantly better performance than alignment-based approaches, BLASTP and PHMMER, in both prediction accuracy and speed. PEPPER is anticipated to facilitate community-wide efforts to conduct high-throughput identification and analysis of PE_PGRS proteins.

A Deep Learning Approach for Predicting Antigenic Variation of Influenza A H3N2

Modeling antigenic variation in influenza (flu) virus A H3N2 using amino acid sequences is a promising approach for improving the prediction accuracy of immune efficacy of vaccines and increasing the efficiency of vaccine screening. Antigenic drift and antigenic jump/shift, which arise from the accumulation of mutations with small or moderate effects and from a major, abrupt change with large effects on the surface antigen hemagglutinin (HA), respectively, are two types of antigenic variation that facilitate immune evasion of flu virus A and make it challenging to predict the antigenic properties of new viral strains. Despite considerable progress in modeling antigenic variation based on the amino acid sequences, few studies focus on the deep learning framework which could be most suitable to be applied to this task. Here, we propose a novel deep learning approach that incorporates a convolutional neural network (CNN) and bidirectional long-short-term memory (BLSTM) neural network to predict antigenic variation. In this approach, CNN extracts the complex local contexts of amino acids while the BLSTM neural network captures the long-distance sequence information. When compared to the existing methods, our deep learning approach achieves the overall highest prediction performance on the validation dataset, and more encouragingly, it achieves prediction agreements of 99.20% and 96.46% for the strains in the forthcoming year and in the next two years included in an existing set of chronological amino acid sequences, respectively. These results indicate that our deep learning approach is promising to be applied to antigenic variation prediction of flu virus A H3N2.

Metapredict: a fast, accurate, and easy-to-use predictor of consensus disorder and structure

Intrinsically disordered proteins and protein regions make up a substantial fraction of many proteomes in which they play a wide variety of essential roles. A critical first step in understanding the role of disordered protein regions in biological function is to identify those disordered regions correctly. Computational methods for disorder prediction have emerged as a core set of tools to guide experiments, interpret results, and develop hypotheses. Given the multiple different predictors available, consensus scores have emerged as a popular approach to mitigate biases or limitations of any single method. Consensus scores integrate the outcome of multiple independent disorder predictors and provide a per-residue value that reflects the number of tools that predict a residue to be disordered. Although consensus scores help mitigate the inherent problems of using any single disorder predictor, they are computationally expensive to generate. They also necessitate the installation of multiple different software tools, which can be prohibitively difficult. To address this challenge, we developed a deep-learning-based predictor of consensus disorder scores. Our predictor, metapredict, utilizes a bidirectional recurrent neural network trained on the consensus disorder scores from 12 proteomes. By benchmarking metapredict using two orthogonal approaches, we found that metapredict is among the most accurate disorder predictors currently available. Metapredict is also remarkably fast, enabling proteome-scale disorder prediction in minutes. Importantly, metapredict is a fully open source and is distributed as a Python package, a collection of command-line tools, and a web server, maximizing the potential practical utility of the predictor. We believe metapredict offers a convenient, accessible, accurate, and high-performance predictor for single-proteins and proteomes alike.

Metapredict: a fast, accurate, and easy-to-use predictor of consensus disorder and structure

Intrinsically disordered proteins and protein regions make up a substantial fraction of many proteomes in which they play a wide variety of essential roles. A critical first step in understanding the role of disordered protein regions in biological function is to identify those disordered regions correctly. Computational methods for disorder prediction have emerged as a core set of tools to guide experiments, interpret results, and develop hypotheses. Given the multiple different predictors available, consensus scores have emerged as a popular approach to mitigate biases or limitations of any single method. Consensus scores integrate the outcome of multiple independent disorder predictors and provide a per-residue value that reflects the number of tools that predict a residue to be disordered. Although consensus scores help mitigate the inherent problems of using any single disorder predictor, they are computationally expensive to generate. They also necessitate the installation of multiple different software tools, which can be prohibitively difficult. To address this challenge, we developed a deep-learning-based predictor of consensus disorder scores. Our predictor, metapredict, utilizes a bidirectional recurrent neural network trained on the consensus disorder scores from 12 proteomes. By benchmarking metapredict using two orthogonal approaches, we found that metapredict is among the most accurate disorder predictors currently available. Metapredict is also remarkably fast, enabling proteome-scale disorder prediction in minutes. Importantly, metapredict is a fully open source and is distributed as a Python package, a collection of command-line tools, and a web server, maximizing the potential practical utility of the predictor. We believe metapredict offers a convenient, accessible, accurate, and high-performance predictor for single-proteins and proteomes alike.

PARROT is a flexible recurrent neural network framework for analysis of large protein datasets

The rise of high-throughput experiments has transformed how scientists approach biological questions. The ubiquity of large-scale assays that can test thousands of samples in a day has necessitated the development of new computational approaches to interpret this data. Among these tools, machine learning approaches are increasingly being utilized due to their ability to infer complex nonlinear patterns from high-dimensional data. Despite their effectiveness, machine learning (and in particular deep learning) approaches are not always accessible or easy to implement for those with limited computational expertise. Here we present PARROT, a general framework for training and applying deep learning-based predictors on large protein datasets. Using an internal recurrent neural network architecture, PARROT is capable of tackling both classification and regression tasks while only requiring raw protein sequences as input. We showcase the potential uses of PARROT on three diverse machine learning tasks: predicting phosphorylation sites, predicting transcriptional activation function of peptides generated by high-throughput reporter assays, and predicting the fibrillization propensity of amyloid beta with data generated by deep mutational scanning. Through these examples, we demonstrate that PARROT is easy to use, performs comparably to state-of-the-art computational tools, and is applicable for a wide array of biological problems.

Remote homology clustering identifies lowly conserved families of effector proteins in plant-pathogenic fungi

Plant diseases caused by fungal pathogens are typically initiated by molecular interactions between 'effector' molecules released by a pathogen and receptor molecules on or within the plant host cell. In many cases these effector-receptor interactions directly determine host resistance or susceptibility. The search for fungal effector proteins is a developing area in fungal-plant pathology, with more than 165 distinct confirmed fungal effector proteins in the public domain. For a small number of these, novel effectors can be rapidly discovered across multiple fungal species through the identification of known effector homologues. However, many have no detectable homology by standard sequence-based search methods. This study employs a novel comparison method (RemEff) that is capable of identifying protein families with greater sensitivity than traditional homology-inference methods, leveraging a growing pool of confirmed fungal effector data to enable the prediction of novel fungal effector candidates by protein family association. Resources relating to the RemEff method and data used in this study are available from https://figshare.com/projects/Effector_protein_remote_homology/87965.

DeepDRBP-2L: A New Genome Annotation Predictor for Identifying DNA-Binding Proteins and RNA-Binding Proteins Using Convolutional Neural Network and Long Short-Term Memory

DNA-binding proteins (DBPs) and RNA-binding proteins (RBPs) are two kinds of crucial proteins, which are associated with various cellule activities and some important diseases. Accurate identification of DBPs and RBPs facilitate both theoretical research and real world application. Existing sequence-based DBP predictors can accurately identify DBPs but incorrectly predict many RBPs as DBPs, and vice versa, resulting in low prediction precision. Moreover, some proteins (DRBPs) interacting with both DNA and RNA play important roles in gene expression and cannot be identified by existing computational methods. In this study, a two-level predictor named DeepDRBP-2L was proposed by combining Convolutional Neural Network (CNN) and the Long Short-Term Memory (LSTM). It is the first computational method that is able to identify DBPs, RBPs and DRBPs. Rigorous cross-validations and independent tests showed that DeepDRBP-2L is able to overcome the shortcoming of the existing methods and can go one further step to identify DRBPs. Application of DeepDRBP-2L to tomato genome further demonstrated its performance. The webserver of DeepDRBP-2L is freely available at http://bliulab.net/DeepDRBP-2L.

Improving protein domain classification for third-generation sequencing reads using deep learning

BACKGROUND

With the development of third-generation sequencing (TGS) technologies, people are able to obtain DNA sequences with lengths from 10s to 100s of kb. These long reads allow protein domain annotation without assembly, thus can produce important insights into the biological functions of the underlying data. However, the high error rate in TGS data raises a new challenge to established domain analysis pipelines. The state-of-the-art methods are not optimized for noisy reads and have shown unsatisfactory accuracy of domain classification in TGS data. New computational methods are still needed to improve the performance of domain prediction in long noisy reads.

RESULTS

In this work, we introduce ProDOMA, a deep learning model that conducts domain classification for TGS reads. It uses deep neural networks with 3-frame translation encoding to learn conserved features from partially correct translations. In addition, we formulate our problem as an open-set problem and thus our model can reject reads not containing the targeted domains. In the experiments on simulated long reads of protein coding sequences and real TGS reads from the human genome, our model outperforms HMMER and DeepFam on protein domain classification.

CONCLUSIONS

In summary, ProDOMA is a useful end-to-end protein domain analysis tool for long noisy reads without relying on error correction.

ATSE: a peptide toxicity predictor by exploiting structural and evolutionary information based on graph neural network and attention mechanism

MOTIVATION

Peptides have recently emerged as promising therapeutic agents against various diseases. For both research and safety regulation purposes, it is of high importance to develop computational methods to accurately predict the potential toxicity of peptides within the vast number of candidate peptides.

RESULTS

In this study, we proposed ATSE, a peptide toxicity predictor by exploiting structural and evolutionary information based on graph neural networks and attention mechanism. More specifically, it consists of four modules: (i) a sequence processing module for converting peptide sequences to molecular graphs and evolutionary profiles, (ii) a feature extraction module designed to learn discriminative features from graph structural information and evolutionary information, (iii) an attention module employed to optimize the features and (iv) an output module determining a peptide as toxic or non-toxic, using optimized features from the attention module.

CONCLUSION

Comparative studies demonstrate that the proposed ATSE significantly outperforms all other competing methods. We found that structural information is complementary to the evolutionary information, effectively improving the predictive performance. Importantly, the data-driven features learned by ATSE can be interpreted and visualized, providing additional information for further analysis. Moreover, we present a user-friendly online computational platform that implements the proposed ATSE, which is now available at http://server.malab.cn/ATSE. We expect that it can be a powerful and useful tool for researchers of interest.

Deep Robust Framework for Protein Function Prediction Using Variable-Length Protein Sequences

The order of amino acids in a protein sequence enables the protein to acquire a conformation suitable for performing functions, thereby motivating the need to analyze these sequences for predicting functions. Although machine learning based approaches are fast compared to methods using BLAST, FASTA, etc., they fail to perform well for long protein sequences (with more than 300 amino acids). In this paper, we introduce a novel method for construction of two separate feature sets for protein using bi-directional long short-term memory network based on the analysis of fixed 1) single-sized segments and 2) multi-sized segments. The model trained on the proposed feature set based on multi-sized segments is combined with the model trained using state-of-the-art Multi-label Linear Discriminant Analysis (MLDA) features to further improve the accuracy. Extensive evaluations using separate datasets for biological processes and molecular functions demonstrate not only improved results for long sequences, but also significantly improve the overall accuracy over state-of-the-art method. The single-sized approach produces an improvement of +3.37 percent for biological processes and +5.48 percent for molecular functions over the MLDA based classifier. The corresponding numbers for multi-sized approach are +5.38 and +8.00 percent. Combining the two models, the accuracy further improves to +7.41 and +9.21 percent, respectively.

DeepTorrent: a deep learning-based approach for predicting DNA N4-methylcytosine sites

DNA N4-methylcytosine (4mC) is an important epigenetic modification that plays a vital role in regulating DNA replication and expression. However, it is challenging to detect 4mC sites through experimental methods, which are time-consuming and costly. Thus, computational tools that can identify 4mC sites would be very useful for understanding the mechanism of this important type of DNA modification. Several machine learning-based 4mC predictors have been proposed in the past 3 years, although their performance is unsatisfactory. Deep learning is a promising technique for the development of more accurate 4mC site predictions. In this work, we propose a deep learning-based approach, called DeepTorrent, for improved prediction of 4mC sites from DNA sequences. It combines four different feature encoding schemes to encode raw DNA sequences and employs multi-layer convolutional neural networks with an inception module integrated with bidirectional long short-term memory to effectively learn the higher-order feature representations. Dimension reduction and concatenated feature maps from the filters of different sizes are then applied to the inception module. In addition, an attention mechanism and transfer learning techniques are also employed to train the robust predictor. Extensive benchmarking experiments demonstrate that DeepTorrent significantly improves the performance of 4mC site prediction compared with several state-of-the-art methods.

UDSMProt: universal deep sequence models for protein classification

MOTIVATION

Inferring the properties of a protein from its amino acid sequence is one of the key problems in bioinformatics. Most state-of-the-art approaches for protein classification are tailored to single classification tasks and rely on handcrafted features, such as position-specific-scoring matrices from expensive database searches. We argue that this level of performance can be reached or even be surpassed by learning a task-agnostic representation once, using self-supervised language modeling, and transferring it to specific tasks by a simple fine-tuning step.

RESULTS

We put forward a universal deep sequence model that is pre-trained on unlabeled protein sequences from Swiss-Prot and fine-tuned on protein classification tasks. We apply it to three prototypical tasks, namely enzyme class prediction, gene ontology prediction and remote homology and fold detection. The proposed method performs on par with state-of-the-art algorithms that were tailored to these specific tasks or, for two out of three tasks, even outperforms them. These results stress the possibility of inferring protein properties from the sequence alone and, on more general grounds, the prospects of modern natural language processing methods in omics. Moreover, we illustrate the prospects for explainable machine learning methods in this field by selected case studies.

AVAILABILITY AND IMPLEMENTATION

Source code is available under https://github.com/nstrodt/UDSMProt.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

BioSeq-Analysis: a platform for DNA, RNA and protein sequence analysis based on machine learning approaches

With the avalanche of biological sequences generated in the post-genomic age, one of the most challenging problems is how to computationally analyze their structures and functions. Machine learning techniques are playing key roles in this field. Typically, predictors based on machine learning techniques contain three main steps: feature extraction, predictor construction and performance evaluation. Although several Web servers and stand-alone tools have been developed to facilitate the biological sequence analysis, they only focus on individual step. In this regard, in this study a powerful Web server called BioSeq-Analysis (http://bioinformatics.hitsz.edu.cn/BioSeq-Analysis/) has been proposed to automatically complete the three main steps for constructing a predictor. The user only needs to upload the benchmark data set. BioSeq-Analysis can generate the optimized predictor based on the benchmark data set, and the performance measures can be reported as well. Furthermore, to maximize user's convenience, its stand-alone program was also released, which can be downloaded from http://bioinformatics.hitsz.edu.cn/BioSeq-Analysis/download/, and can be directly run on Windows, Linux and UNIX. Applied to three sequence analysis tasks, experimental results showed that the predictors generated by BioSeq-Analysis even outperformed some state-of-the-art methods. It is anticipated that BioSeq-Analysis will become a useful tool for biological sequence analysis.

Critiquing Protein Family Classification Models Using Sufficient Input Subsets

In many application domains, neural networks are highly accurate and have been deployed at large scale. However, users often do not have good tools for understanding how these models arrive at their predictions. This has hindered adoption in fields such as the life and medical sciences, where researchers require that models base their decisions on underlying biological phenomena rather than peculiarities of the dataset. We propose a set of methods for critiquing deep learning models and demonstrate their application for protein family classification, a task for which high-accuracy models have considerable potential impact. Our methods extend the Sufficient Input Subsets (SIS) technique, which we use to identify subsets of features in each protein sequence that are alone sufficient for classification. Our suite of tools analyzes these subsets to shed light on the decision-making criteria employed by models trained on this task. These tools show that while deep models may perform classification for biologically relevant reasons, their behavior varies considerably across the choice of network architecture and parameter initialization. While the techniques that we develop are specific to the protein sequence classification task, the approach taken generalizes to a broad set of scientific contexts in which model interpretability is essential.

Deep learning for mining protein data

The recent emergence of deep learning to characterize complex patterns of protein big data reveals its potential to address the classic challenges in the field of protein data mining. Much research has revealed the promise of deep learning as a powerful tool to transform protein big data into valuable knowledge, leading to scientific discoveries and practical solutions. In this review, we summarize recent publications on deep learning predictive approaches in the field of mining protein data. The application architectures of these methods include multilayer perceptrons, stacked autoencoders, deep belief networks, two- or three-dimensional convolutional neural networks, recurrent neural networks, graph neural networks, and complex neural networks and are described from five perspectives: residue-level prediction, sequence-level prediction, three-dimensional structural analysis, interaction prediction, and mass spectrometry data mining. The advantages and deficiencies of these architectures are presented in relation to various tasks in protein data mining. Additionally, some practical issues and their future directions are discussed, such as robust deep learning for protein noisy data, architecture optimization for specific tasks, efficient deep learning for limited protein data, multimodal deep learning for heterogeneous protein data, and interpretable deep learning for protein understanding. This review provides comprehensive perspectives on general deep learning techniques for protein data analysis.

DeepSVM-fold: protein fold recognition by combining support vector machines and pairwise sequence similarity scores generated by deep learning networks

Protein fold recognition is critical for studying the structures and functions of proteins. The existing protein fold recognition approaches failed to efficiently calculate the pairwise sequence similarity scores of the proteins in the same fold sharing low sequence similarities. Furthermore, the existing feature vectorization strategies are not able to measure the global relationships among proteins from different protein folds. In this article, we proposed a new computational predictor called DeepSVM-fold for protein fold recognition by introducing a new feature vector based on the pairwise sequence similarity scores calculated from the fold-specific features extracted by deep learning networks. The feature vectors are then fed into a support vector machine to construct the predictor. Experimental results on the benchmark dataset (LE) show that DeepSVM-fold obviously outperforms all the other competing methods.

Unified rational protein engineering with sequence-based deep representation learning

Rational protein engineering requires a holistic understanding of protein function. Here, we apply deep learning to unlabeled amino-acid sequences to distill the fundamental features of a protein into a statistical representation that is semantically rich and structurally, evolutionarily and biophysically grounded. We show that the simplest models built on top of this unified representation (UniRep) are broadly applicable and generalize to unseen regions of sequence space. Our data-driven approach predicts the stability of natural and de novo designed proteins, and the quantitative function of molecularly diverse mutants, competitively with the state-of-the-art methods. UniRep further enables two orders of magnitude efficiency improvement in a protein engineering task. UniRep is a versatile summary of fundamental protein features that can be applied across protein engineering informatics.

Sparse Convolutional Denoising Autoencoders for Genotype Imputation

Genotype imputation, where missing genotypes can be computationally imputed, is an essential tool in genomic analysis ranging from genome wide associations to phenotype prediction. Traditional genotype imputation methods are typically based on haplotype-clustering algorithms, hidden Markov models (HMMs), and statistical inference. Deep learning-based methods have been recently reported to suitably address the missing data problems in various fields. To explore the performance of deep learning for genotype imputation, in this study, we propose a deep model called a sparse convolutional denoising autoencoder (SCDA) to impute missing genotypes. We constructed the SCDA model using a convolutional layer that can extract various correlation or linkage patterns in the genotype data and applying a sparse weight matrix resulted from the L₁ regularization to handle high dimensional data. We comprehensively evaluated the performance of the SCDA model in different scenarios for genotype imputation on the yeast and human genotype data, respectively. Our results showed that SCDA has strong robustness and significantly outperforms popular reference-free imputation methods. This study thus points to another novel application of deep learning models for missing data imputation in genomic studies.

ProtDet-CCH: Protein Remote Homology Detection by Combining Long Short-Term Memory and Ranking Methods

As one of the most challenging tasks in sequence analysis, protein remote homology detection has been extensively studied. Methods based on discriminative models and ranking approaches have achieved the state-of-the-art performance, and these two kinds of methods are complementary. In this study, three LSTM models have been applied to construct the predictors for protein remote homology detection, including ULSTM, BLSTM, and CNN-BLSTM. They are able to automatically extract the local and global sequence order information. Combined with PSSMs, the CNN-BLSTM achieved the best performance among the three LSTM-based models. We named this method as CNN-BLSTM-PSSM. Finally, a new method called ProtDet-CCH was proposed by combining CNN-BLSTM-PSSM and a ranking method HHblits. Tested on a widely used SCOP benchmark dataset, ProtDet-CCH achieved an ROC score of 0.998, and an ROC50 score of 0.982, significantly outperforming other existing state-of-the-art methods. Experimental results on two updated SCOPe independent datasets showed that ProtDet-CCH can achieve stable performance. Furthermore, our method can provide useful insights for studying the features and motifs of protein families and superfamilies. It is anticipated that ProtDet-CCH will become a very useful tool for protein remote homology detection.

Transcription factors–DNA interactions in rice: identification and verification

The completion of the rice genome sequence paved the way for rice functional genomics research. Additionally, the functional characterization of transcription factors is currently a popular and crucial objective among researchers. Transcription factors are one of the groups of proteins that bind to either enhancer or promoter regions of genes to regulate expression. On the basis of several typical examples of transcription factor analyses, we herein summarize selected research strategies and methods and introduce their advantages and disadvantages. This review may provide some theoretical and technical guidelines for future investigations of transcription factors, which may be helpful to develop new rice varieties with ideal traits.

MinE-RFE: determine the optimal subset from RFE by minimizing the subset-accuracy–defined energy

Recursive feature elimination (RFE), as one of the most popular feature selection algorithms, has been extensively applied to bioinformatics. During the training, a group of candidate subsets are generated by iteratively eliminating the least important features from the original features. However, how to determine the optimal subset from them still remains ambiguous. Among most current studies, either overall accuracy or subset size (SS) is used to select the most predictive features. Using which one or both and how they affect the prediction performance are still open questions. In this study, we proposed MinE-RFE, a novel RFE-based feature selection approach by sufficiently considering the effect of both factors. Subset decision problem was reflected into subset-accuracy space and became an energy-minimization problem. We also provided a mathematical description of the relationship between the overall accuracy and SS using Gaussian Mixture Models together with spline fitting. Besides, we comprehensively reviewed a variety of state-of-the-art applications in bioinformatics using RFE. We compared their approaches of deciding the final subset from all the candidate subsets with MinE-RFE on diverse bioinformatics data sets. Additionally, we also compared MinE-RFE with some well-used feature selection algorithms. The comparative results demonstrate that the proposed approach exhibits the best performance among all the approaches. To facilitate the use of MinE-RFE, we further established a user-friendly web server with the implementation of the proposed approach, which is accessible at http://qgking.wicp.net/MinE/. We expect this web server will be a useful tool for research community.

A Review of DNA-binding Proteins Prediction Methods

Background:

DNA-binding proteins, binding to DNA, widely exist in living cells, participating in many cell activities. They can participate some DNA-related cell activities, for instance DNA replication, transcription, recombination, and DNA repair.

Objective:

Given the importance of DNA-binding proteins, studies for predicting the DNA-binding proteins have been a popular issue over the past decades. In this article, we review current machine-learning methods which research on the prediction of DNA-binding proteins through feature representation methods, classifiers, measurements, dataset and existing web server.

Method:

The prediction methods of DNA-binding protein can be divided into two types, based on amino acid composition and based on protein structure. In this article, we accord to the two types methods to introduce the application of machine learning in DNA-binding proteins prediction.

Results:

Machine learning plays an important role in the classification of DNA-binding proteins, and the result is better. The best ACC is above 80%.

Conclusion:

Machine learning can be widely used in many aspects of biological information, especially in protein classification. Some issues should be considered in future work. First, the relationship between the number of features and performance must be explored. Second, many features are used to predict DNA-binding proteins and propose solutions for high-dimensional spaces.

Gene2vec: gene subsequence embedding for prediction of mammalian N6-methyladenosine sites from mRNA

N⁶-Methyladenosine (m⁶A) refers to methylation modification of the adenosine nucleotide acid at the nitrogen-6 position. Many conventional computational methods for identifying N⁶-methyladenosine sites are limited by the small amount of data available. Taking advantage of the thousands of m⁶A sites detected by high-throughput sequencing, it is now possible to discover the characteristics of m⁶A sequences using deep learning techniques. To the best of our knowledge, our work is the first attempt to use word embedding and deep neural networks for m⁶A prediction from mRNA sequences. Using four deep neural networks, we developed a model inferred from a larger sequence shifting window that can predict m⁶A accurately and robustly. Four prediction schemes were built with various RNA sequence representations and optimized convolutional neural networks. The soft voting results from the four deep networks were shown to outperform all of the state-of-the-art methods. We evaluated these predictors mentioned above on a rigorous independent test data set and proved that our proposed method outperforms the state-of-the-art predictors. The training, independent, and cross-species testing data sets are much larger than in previous studies, which could help to avoid the problem of overfitting. Furthermore, an online prediction web server implementing the four proposed predictors has been built and is available at http://server.malab.cn/Gene2vec/.

Protein fold recognition based on multi-view modeling

Motivation

Protein fold recognition has attracted increasing attention because it is critical for studies of the 3D structures of proteins and drug design. Researchers have been extensively studying this important task, and several features with high discriminative power have been proposed. However, the development of methods that efficiently combine these features to improve the predictive performance remains a challenging problem.

Results

In this study, we proposed two algorithms: MV-fold and MT-fold. MV-fold is a new computational predictor based on the multi-view learning model for fold recognition. Different features of proteins were treated as different views of proteins, including the evolutionary information, secondary structure information and physicochemical properties. These different views constituted the latent space. The ε-dragging technique was employed to enlarge the margins between different protein folds, improving the predictive performance of MV-fold. Then, MV-fold was combined with two template-based methods: HHblits and HMMER. The ensemble method is called MT-fold incorporating the advantages of both discriminative methods and template-based methods. Experimental results on five widely used benchmark datasets (DD, RDD, EDD, TG and LE) showed that the proposed methods outperformed some state-of-the-art methods in this field, indicating that MV-fold and MT-fold are useful computational tools for protein fold recognition and protein homology detection and would be efficient tools for protein sequence analysis. Finally, we constructed an update and rigorous benchmark dataset based on SCOPe (version 2.07) to fairly evaluate the performance of the proposed method, and our method achieved stable performance on this new dataset. This new benchmark dataset will become a widely used benchmark dataset to fairly evaluate the performance of different methods for fold recognition.

Supplementary information

Supplementary data are available at Bioinformatics online.

HITS-PR-HHblits: protein remote homology detection by combining PageRank and Hyperlink-Induced Topic Search

As one of the most important fundamental problems in protein sequence analysis, protein remote homology detection is critical for both theoretical research (protein structure and function studies) and real world applications (drug design). Although several computational predictors have been proposed, their detection performance is still limited. In this study, we treat protein remote homology detection as a document retrieval task, where the proteins are considered as documents and its aim is to find the highly related documents with the query documents in a database. A protein similarity network was constructed based on the true labels of proteins in the database, and the query proteins were then connected into the network based on the similarity scores calculated by three ranking methods, including PSI-BLAST, Hmmer and HHblits. The PageRank algorithm and Hyperlink-Induced Topic Search (HITS) algorithm were respectively performed on this network to move the homologous proteins of query proteins to the neighbors of the query proteins in the network. Finally, PageRank and HITS algorithms were combined, and a predictor called HITS-PR-HHblits was proposed to further improve the predictive performance. Tested on the SCOP and SCOPe benchmark datasets, the experimental results showed that the proposed protocols outperformed other state-of-the-art methods. For the convenience of the most experimental scientists, a web server for HITS-PR-HHblits was established at http://bioinformatics.hitsz.edu.cn/HITS-PR-HHblits, by which the users can easily get the results without the need to go through the mathematical details. The HITS-PR-HHblits predictor is a protocol for protein remote homology detection using different sets of programs, which will become a very useful computational tool for proteome analysis.

Rise of Deep Learning for Genomic, Proteomic, and Metabolomic Data Integration in Precision Medicine

Machine learning (ML) is being ubiquitously incorporated into everyday products such as Internet search, email spam filters, product recommendations, image classification, and speech recognition. New approaches for highly integrated manufacturing and automation such as the Industry 4.0 and the Internet of things are also converging with ML methodologies. Many approaches incorporate complex artificial neural network architectures and are collectively referred to as deep learning (DL) applications. These methods have been shown capable of representing and learning predictable relationships in many diverse forms of data and hold promise for transforming the future of omics research and applications in precision medicine. Omics and electronic health record data pose considerable challenges for DL. This is due to many factors such as low signal to noise, analytical variance, and complex data integration requirements. However, DL models have already been shown capable of both improving the ease of data encoding and predictive model performance over alternative approaches. It may not be surprising that concepts encountered in DL share similarities with those observed in biological message relay systems such as gene, protein, and metabolite networks. This expert review examines the challenges and opportunities for DL at a systems and biological scale for a precision medicine readership.

Taxonomic Classification for Living Organisms Using Convolutional Neural Networks

Taxonomic classification has a wide-range of applications such as finding out more about evolutionary history. Compared to the estimated number of organisms that nature harbors, humanity does not have a thorough comprehension of to which specific classes they belong. The classification of living organisms can be done in many machine learning techniques. However, in this study, this is performed using convolutional neural networks. Moreover, a DNA encoding technique is incorporated in the algorithm to increase performance and avoid misclassifications. The algorithm proposed outperformed the state of the art algorithms in terms of accuracy and sensitivity, which illustrates a high potential for using it in many other applications in genome analysis.