Search

Chromatin immunoprecipitation (ChIP-Seq) has emerged as a superior alternative to microarray technology as it provides higher resolution, less noise, greater coverage and wider dynamic range. While ChIP-Seq enables probing of DNA-protein interaction over the entire genome, it requires the use of sophisticated tools to recognize hidden patterns and extract meaningful data. Over the years, various attempts have resulted in several algorithms making use of different heuristics to accurately determine individual peaks corresponding to unique DNA-protein. However, finding all the significant peaks with high accuracy in a reasonable time is still a challenge. In this work, we propose the use of Multi-level thresholding algorithm, which we call LinMLTBS, used to identify the enriched regions on ChIP-Seq data. Although various suboptimal heuristics have been proposed for multi-level thresholding, we emphasize on the use of an algorithm capable of obtaining an optimal solution, while maintaining linear-time complexity. Testing various algorithm on various ENCODE project datasets shows that our approach attains higher accuracy relative to previously proposed peak finders while retaining a reasonable processing speed.

( Early Access ) Drug combination therapies have been a viable strategy for the treatment of complex diseases such as cancer due to increased efficacy and reduced side effects. However, experimentally validating all possible combinations for synergistic interaction even with high-throughout screens is intractable due to vast combinatorial search space. Computational techniques can reduce the number of combinations to be evaluated experimentally by prioritizing promising candidates. We present MatchMaker that predicts drug synergy scores using drug chemical structure information and gene expression profiles of cell lines in a deep learning framework. For the first time, our model utilizes the largest known drug combination dataset to date, DrugComb. We compare the performance of MatchMaker with the state-of-the-art models and observe up to 15% correlation and 33% mean squared error (MSE) improvements over the next best method. We investigate the cell types and drug pairs that are relatively harder to predict and present novel candidate pairs. MatchMaker is built and available at https://github.com/tastanlab/matchmaker

It has been proved that long noncoding RNA (lncRNA) plays critical roles in many human diseases. Therefore, inferring associations between lncRNAs and diseases can contribute to disease diagnosis, prognosis and treatment. To overcome the limitation of traditional experimental methods such as expensive and time-consuming, several computational methods have been proposed to predict lncRNA-disease associations by fusing different biological data. However, the prediction performance of lncRNA-disease associations identification needs to be improved. In this study, we propose a computational model (named LDICDL) to identify lncRNA-disease associations based on collaborative deep learning. It uses an automatic encoder to denoise multiple lncRNA feature information and multiple disease feature information, respectively. Then, the matrix decomposition algorithm is employed to predict the potential lncRNA-disease associations. In addition, to overcome the limitation of matrix decomposition, the hybrid model is developed to predict associations between new lncRNA (or disease) and diseases (or lncRNA). The ten-fold cross validation and de novo test are applied to evaluate the performance of method. The experimental results show LDICDL outperforms than other state-of-the-art methods in prediction performance.

The opioid abuse epidemic represents a major public health threat to global populations. The role social media may play in facilitating illicit drug trade is largely unknown due to limited research. However, it is known that social media use among adults in the US is widespread, there is vast capability for online promotion of illegal drugs with delayed or limited deterrence of such messaging, and further, general commercial sale applications provide safeguards for transactions; however, they do not discriminate between legal and illegal sale transactions. These characteristics of the social media environment present challenges to surveillance which is needed for advancing knowledge of online drug markets and the role they play in the drug abuse and overdose deaths. In this paper, we present a computational framework developed to automatically detect illicit drug ads and communities of vendors. The SVM- and CNN- based methods for detecting illicit drug ads, and a matrix factorization based method for discovering overlapping communities have been extensively validated on the large dataset collected from Google+, Flickr and Tumblr. Pilot test results demonstrate that our computational methods can effectively identify illicit drug ads and detect vendor-community with accuracy. These methods hold promise to advance scientific knowledge surrounding the role social media may play in perpetuating the drug abuse epidemic.

An ordinary differential equation (ODE) model of the working of the thyroid system for euthyroidism has been presented. As clinical data for thyroid hormones is relatively scarce, such modelling offers potential benefits over wet lab procedures. Genetic algorithms developed for determining of parameters of the ODE system using the available data have been presented and evaluated. This approach enables subject specific parameter estimation towards characterisation of individual thyroid operation. Initially, a simple steady state model was used. Later a cosinor model for the circadian variation of thyroid hormones was used to obtain more reliable results, as indicated through sensitivity analysis in conjunction with other statistical methods. Our parameter determination method has been tested on groups of patients with similar observed values of thyroid stimulating hormone (TSH), free T3 and free T4 (identified through clustering) to determine their parameter values jointly. This approach appears to produce parameter sets with lower variation than parameters determined independently, thus leading to better parameter determination.

The inverse Potts problem to infer a Boltzmann distribution for homologous protein sequences from their single-site and pairwise amino acid frequencies recently attracts a great deal of attention in the studies of protein structure and evolution. We study regularization and learning methods and how to tune regularization parameters to correctly infer interactions in Boltzmann machine learning. Using $L_2$ regularization for fields, group $L_1$ for couplings is shown to be very effective for sparse couplings in comparison with $L_2$ and $L_1$. Two regularization parameters are tuned to yield equal values for both the sample and ensemble averages of evolutionary energy. Both averages smoothly change and converge, but their learning profiles are very different between learning methods. The Adam method is modified to make stepsize proportional to the gradient for sparse couplings and to use a soft-thresholding function for group $L_1$. It is shown by first inferring interactions from protein sequences and then from Monte Carlo samples that the fields and couplings can be well recovered, but that recovering the pairwise correlations in the resolution of a total energy is harder for the natural proteins than for the protein-like sequences. Selective temperature for folding/structural constrains in protein evolution is also estimated., In this version (arXiv:1909.05006v3), the values of selective temperature for protein PF00153, $T_s$ in Table 5 and in the section 2.8, and folding free energy for PF00595 are corrected and shown in red. The version 2 (arXiv:1909.05006v2) was published in the IEEE/ACM Transactions on Computational Biology and Bioinformatics. The program is available from https://gitlab.com/sanzo.miyazawa/BM/

A drug-drug interaction (DDI) is defined as an association between two drugs where the pharmacological effects of a drug are influenced by another drug. Positive DDIs can usually improve the therapeutic effects of patients, but negative DDIs cause the major cause of adverse drug reactions and even result in the drug withdrawal from the market and the patient death. Therefore, identifying DDIs has become a key component of the drug development and disease treatment. In this study, we propose a novel method to predict DDIs based on the integrated similarity and semi-supervised learning (DDI-IS-SL). DDI-IS-SL integrates the drug chemical, biological and phenotype data to calculate the feature similarity of drugs with the cosine similarity method. The Gaussian Interaction Profile kernel similarity of drugs is also calculated based on known DDIs. A semi-supervised learning method (the Regularized Least Squares classifier) is used to calculate the interaction possibility scores of drug-drug pairs. In terms of the 5-fold cross validation, 10-fold cross validation and de novo drug validation, DDI-IS-SL can achieve the better prediction performance than other comparative methods. In addition, the average computation time of DDI-IS-SL is shorter than that of other comparative methods. Finally, case studies further demonstrate the performance of DDI-IS-SL in practical applications.

Non-small cell lung cancer with an activating epidermal growth factor receptor (EGFR) mutation responds well to targeted drugs. In most cases, drug resistance appears after about a year. Several studies have been conducted on the kinase domain of EGFR to understand the drug resistance mechanism. Since EGFR is a multi-domain protein, mutation in the kinase domain may affect the other domains as well. In this study, we examine the complete structure of the multi-domain EGFR protein and its mutants. We performed molecular dynamics simulations for wildtype EGFR, EGFR with L858R mutation, and EGFR with L858R and T790M mutations. We applied normal mode analysis and complex network analysis to extract the correlated motions in the domains of EGFR. The normal modes are used to construct the dynamic cross-correlation map (DCCM). Simulation results show different patterns of correlated motions in each domain of EGFR mutants compared to the wildtype. In Domains 1 and 3 of the extracellular region, a small number of weak positively correlated motions are extracted. Domains 2 and 4 show large numbers of both positive and negative motions. However, the negatively correlated motions are stronger in mutant structures compared to the wildtype. In Domain 7, some residues showed a positive correlation around the main diagonal. We also identified different communities, nodes and crucial residues in the domains of the structures, which can be important for the function of EGFR. Moreover, hydrogen bond analysis is performed for the stability analysis. The mutant structures have fewer hydrogen bonds compared to the wildtype. Overall, these findings are useful for understanding the dynamics and communications in EGFR domains.

Highly portable Oxford Nanopore sequencer producing long reads in real-time at low cost has made many breakthroughs in genomics studies. However, a major limitation of nanopore sequencing is its high errors when deciphering DNA sequences from noisy and complex raw data. In this paper, we developed an end-to-end basecaller, SACall, based on convolution layers, transformer self-attention layers and a CTC decoder. In SACall, the convolution layers are used to downsample the signals and capture the local patterns. To achieve the contextual relevance of signals, self-attention layers are adopted to calculate the similarity of the signals at any two positions in the raw signal sequence. Finally, the CTC decoder generates the DNA sequence by a beam search algorithm. We use a benchmark consisting of nine isolated genomes to test the quality of different basecallers including SACall, Albacore, and Guppy. The performances of basecallers are evaluated from the perspective of read accuracy, assembly quality, and consensus accuracy. Among most of the genomes in the test benchmark, the reads basecalled by SACall have fewer errors than the reads basecalled by other basecallers. When assembling the basecalled reads of each genome, the assembly from SACall basecalled reads achieves a higher assembly identity. In addition, there are fewer errors in the polished assembly from reads basecalled by SACall compared to those basecalled by Albacore and Guppy. In general, SACall outperforms the Nanopore official basecallers Albacore and Guppy in the benchmark. Moreover, SACall is an open-source and freely available basecaller, which gives a chance for researchers to train their own basecalling models on specific data and basecall Nanopore reads.

The use of fuzzy logic in machine learning is becoming widespread. In machine learning problems, the data, which have different characteristics, are trained and predicted together. Training the model consisting of data with different characteristics can increase the rate of error in prediction. In this study, we suggest a new approach to assembling prediction with fuzzy clustering. Our approach aims to cluster the data according to their fuzzy membership value and model it with similar characteristics. This approach allows for efficient clustering of objects with more than one cluster characteristic. On the other hand, our approach will enable us to combine boosting type ensemble algorithms, which are various forms of assemblies that are widely used in machine learning due to their excellent success in the literature. We used a mobile game’s customers’ marketing and gameplay data for predicting their customer lifetime value for testing our approach. Customer lifetime value prediction for users is crucial for determining the marketing cost cap for companies. The findings reveal that using a fuzzy method to ensemble the algorithms outperforms implementing the algorithms individually.

In 2013, we released the Structural Antibody Database (SAbDab), a publicly available repository of experimentally determined antibody structures. In the interim, the rapid increase in the number of antibody structure depositions to the Protein Data Bank, driven primarily by increased interest in antibodies as biotherapeutics, has led us to implement several improvements to the original database infrastructure. These include the development of SAbDab-nano, a sub-database that tracks nanobodies (heavy chain-only antibodies) which have seen a particular growth in attention from both the academic and pharmaceutical research communities over the past few years. Both SAbDab and SAbDab-nano are updated weekly, comprehensively annotated with the latest features described here, and are freely accessible at opig.stats.ox.ac.uk/webapps/newsabdab/.

Boolean networks are a notable model of gene regulatory networks and, particularly, prominent theories discuss how they can capture cellular differentiation processes. One frequent motif in gene regulatory networks, especially in those circuits involved in cell differentiation, is autoregulation. In spite of this, the impact of autoregulation on Boolean network attractor landscape has not yet been extensively discussed in literature. In this paper we propose to model autoregulation as self-loops, and analyse how the number of attractors and their robustness may change once they are introduced in a well-known and widely used Boolean networks model, namely random Boolean networks. Results show that self-loops provide an evolutionary advantage in dynamic mechanisms of cells, by increasing both number and maximal robustness of attractors. These results provide evidence to the hypothesis that autoregulation is a straightforward functional component to consolidate cell dynamics, mainly in differentiation processes.

Identifying interactions between drugs and target proteins is a critical step in the drug development process, as it helps identify new targets for drugs and accelerate drug development. The number of known drug-protein interactions (positive samples) is much lower than that of the unknown ones (negative samples), which forms a class imbalance. Most previous methods only utilised part of the negative samples to train the prediction model, so most of the information on negative samples was neglected. Therefore, a new method must be developed to predict candidate drug-related proteins and fully utilise negative samples to improve prediction performance. We present a method based on non-negative matrix factorisation and gradient boosting decision tree (GBDT), named NGDTP, to identify the candidate drug-protein interactions. NGDTP integrates multiple kinds of protein similarities, drugs-proteins interactions, and multiple kinds of drugs similarities at different levels, including target proteins of drugs, drug-related diseases, and side effects of drugs. We propose a network representation learning method based on matrix factorisation to learn low-dimensional vector representations of drug and protein nodes. On the basis of these low-dimensional node representations, a GBDT-based prediction model was constructed and it obtains the association scores through establishing multiple decision trees for a drug-protein pairs. NGDTP is an ensemble learning model that fully utilises all the negative samples to effectively alleviate the problem of class imbalance. NGDTP achieves superior prediction performance when it is compared with several state-of-the-art methods. The experimental results indicate that NGDTP also retrieves more actual drug-protein interactions in the top part of prediction result, which drew significant attention from the biologists. In addition, case studies on 10 drugs further confirmed the ability of the NGDTP to identify potential candidate proteins for drugs.

Mild Cognitive Impairment (MCI) is a preclinical stage of Alzheimer's Disease (AD) and is clinical heterogeneity. The classification of MCI is crucial for the early diagnosis and treatment of AD. In this study, we investigated the potential of using both labeled and unlabeled samples from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort to classify MCI through the multimodal co-training method. We utilized both structural magnetic resonance imaging (sMRI) data and genotype data of 364 MCI samples including 228 labeled and 136 unlabeled MCI samples from the ADNI-1 cohort. First, the selected quantitative trait (QT) features from sMRI data and SNP features from genotype data were used to build two initial classifiers on 228 labeled MCI samples. Then, the co-training method was implemented to obtain new labeled samples from 136 unlabeled MCI samples. Finally, the random forest algorithm was used to obtain a combined classifier to classify MCI patients in the independent ADNI-2 dataset. The experimental results showed that our proposed framework obtains an accuracy of 85.50 percent and an AUC of 0.825 for MCI classification, respectively, which showed that the combined utilization of sMRI and SNP data through the co-training method could significantly improve the performances of MCI classification.

Gene trees can differ from species trees due to a variety of biological phenomena, the most prevalent being gene duplication, horizontal gene transfer, gene loss, and coalescence. To explain topological incongruence between the two trees, researchers apply reconciliation methods, often relying on a maximum parsimony framework. However, while several studies have investigated the space of maximum parsimony reconciliations (MPRs) under the duplication-loss and duplication-transfer-loss models, the space of MPRs under the duplication-loss-coalescence (DLC) model remains poorly understood. To address this problem, we present new algorithms for computing the size of MPR space under the DLC model and sampling from this space uniformly at random. Our algorithms are efficient in practice, with runtime polynomial in the size of the species and gene tree when the number of genes that map to any given species is fixed, thus proving that the MPR problem is fixed-parameter tractable. We have applied our methods to a biological data set of 16 fungal species to provide the first key insights in the space of MPRs under the DLC model. Our results show that a plurality reconciliation, and underlying events, are likely to be representative of MPR space.

Small molecule(SM) drugs can affect the expression of miRNAs, which plays crucial roles in many important biological processes. The chemical structure and clinical information of small molecule can simultaneously incorporate information such as anatomical distribution, therapeutic effects and structural characteristics. It is necessary to develop a novel model that incorporates small molecule chemical structure and clinical information to reveal the unknown small molecule-miRNA associations. In this study, we developed a new framework based on non-negative matrix factorization, called SMANMF, to discover the potential small molecules-miRNAs associations. First, the functional similarity of two miRNAs can be obtained by computing the overlap of the target gene sets in which the miRNAs interact together, and we integrated two types of small molecule similarities, including chemical similarity and clinical similarity. Then, we utilized a non-negative matrix factorization model to discover the unknown relationship between small molecules and miRNAs. The evaluation results indicate that our model can achieve superior prediction performance compared with previous approaches in 5-fold cross-validation. At the same time, the results of case studies also reveal that the SMANMF model has good predictive performance for predicting the potential association between small molecules and miRNAs.

Flavin mono-nucleotides (FMNs) are cofactors that hold responsibility for carrying and transferring electrons in the electron transport chain stage of cellular respiration. Without being facilitated by FMNs, energy production is stagnant due to the interruption in most of the cellular processes. Investigation on FMN's functions, therefore, can gain holistic understanding about human diseases and molecular information on drug targets. We proposed a deep learning model using a two-dimensional convolutional neural network and position specific scoring matrices that could identify FMN interacting residues with the sensitivity of 83.7 percent, specificity of 99.2 percent, accuracy of 98.2 percent, and Matthews correlation coefficients of 0.85 for an independent dataset containing 141 FMN binding sites and 1,920 non-FMN binding sites. The proposed method outperformed other previous studies using similar evaluation metrics. Our positive outcome can also promote the utilization of deep learning in dealing with various problems in bioinformatics and computational biology.

Non-negative matrix factorization (NMF) is a dimensionality reduction technique based on high-dimensional mapping. It can effectively learn part-based representations. In this paper, we propose a method called Dual Hyper-graph Regularized Supervised Non-negative Matrix Factorization (HSNMF). To encode the geometric information of the data, the hyper-graph is introduced into the model as a regularization term. The advantage of hyper-graph learning is to find higher order data relationship to enhance data relevance. This method constructs the data hyper-graph and the feature hyper-graph to find the data manifold and the feature manifold simultaneously. The application of hyper-graph theory in cancer datasets can effectively find pathogenic genes. The discrimination information is further introduced into the objective function to obtain more information about the data. Supervised learning with label information greatly improves the classification effect. Furthermore, the real datasets of cancer usually contain sparse noise, so the -norm is applied to enhance the robustness of HSNMF algorithm. Experiments under The Cancer Genome Atlas (TCGA) datasets verify the feasibility of the HSNMF method.

Gene order evolution of unichromosomal genomes, for example mitochondrial genomes, has been modelled mostly by four major types of genome rearrangements: inversions, transpositions, inverse transpositions, and tandem duplication random losses. Generalizing models that include all those rearrangements while admitting computational tractability are rare. In this paper, we study such a rearrangement model, namely the inverse tandem duplication random loss (iTDRL) model, where an iTDRL duplicates and inverts a continuous segment of a gene order followed by the random loss of one of the redundant copies of each gene. The iTDRL rearrangement has currently been proposed by several authors suggesting it to be a possible mechanisms of mitochondrial gene order evolution. We initiate the algorithmic study of this new model of genome rearrangement by proving that a shortest rearrangement scenario that transforms one given gene order into another given gene order can be obtained in quasilinear time. Furthermore, we show that the length of such a scenario, i.e., the minimum number of iTDRLs in the transformation, can be computed in linear time.

RNA-binding proteins (RBPs) have a significant role in various regulatory tasks. However, the mechanism by which RBPs identify the subsequence target RNAs is still not clear. In recent years, several machine and deep learning-based computational models have been proposed for understanding the binding preferences of RBPs. These methods required integrating multiple features with raw RNA sequences such as secondary structure and their performances can be further improved. In this paper, we propose an efficient and simple convolution neural network, RBPCNN, that relies on the combination of the raw RNA sequence and evolutionary information. We show that conservation scores (evolutionary information) for the RNA sequences can significantly improve the overall performance of the proposed predictor. In addition, the automatic extraction of the binding sequence motifs can enhance our understanding of the binding specificities of RBPs. The experimental results show that RBPCNN outperforms significantly the current state-of-the-art methods. More specifically, the average area under the receiver operator curve was improved by 2.67 percent and the mean average precision was improved by 8.03 percent. The datasets and results can be downloaded from https://home.jbnu.ac.kr/NSCL/RBPCNN.htm.

In recent years, great efforts have been made to analyse biological systems to understand the long-run behaviours. As a well-established formalism for modelling real-life biological systems, Boolean networks (BNs) allow their representation and analysis using formal reasoning and tools. Most biological systems are robust - they can withstand the loss of links and cope with external environmental perturbations. Hence, the BNs used to model such systems are necessarily large and dense, and yet modular. However, existing analysis methods only work well on networks of moderate size. Thus, there is a great need for efficient methods that can handle large-scale BNs and for doing so it is inevitable to exploit both the structural and dynamic properties of the networks. In this paper, we propose a method towards the optimal decomposition of BNs to balance the relation between the structure and dynamics of a network. We show that our method can greatly improve the existing decomposition-based attractor detection by analysing a number of large real-life biological networks.

In the last few years, accumulating evidences had demonstrated that long non-coding RNAs (lncRNAs) participated in the regulation of target gene expression and played an important role in biological processes and human disease development. Thus, prediction of the associations between lncRNAs and disease had become a hot research in the fields of human sophisticated diseases. Most of these methods considered the information of two networks (lncRNA, disease) while neglected other networks. In this study, we designed a multi-layer network by integrating the similarity networks of lncRNAs, diseases and genes, and the known association networks of lncRNA-disease, lncRNAs-gene, and disease-gene, and then we developed a model called MHRWR for predicting the lncRNA-disease potential associations based on random walk with restart. The performance of MHRWR was evaluated by experimentally verified lncRNA-disease associations based on leave-one-out cross validation. MHRWR obtained a reliable AUC value of 0.91344, which significantly outperformed some previous methods. To further validate the reproducibility of performance, we used the model of MHRWR to verify related lncRNAs of colon cancer, colorectal cancer and lung adenocarcinoma in the case studies. The codes of MHRWR is available on: https://github.com/yangyq505/MHRWR.

Single nucleotide variant (SNV) plays an important role in cellular proliferation and tumorigenesis in various types of human cancer. Next-generation sequencing (NGS) has provided high-throughput data at an unprecedented resolution to predict SNVs. Currently, there exist many computational methods for either germline or somatic SNV discovery from NGS data, but very few of them are versatile enough to adapt to any situations. In the absence of matched normal samples, the prediction of somatic SNVs from single-tumor samples becomes considerably challenging, especially when the tumor purity is unknown. Here, we propose a new approach, STIC, to predict somatic SNVs and estimate tumor purity from NGS data without matched normal samples. The main features of STIC include: (1) extracting a set of SNV-relevant features on each site and training the BP neural network algorithm on the features to predict SNVs; (2) creating an iterative process to distinguish somatic SNVs from germline ones by disturbing allele frequency; and (3) establishing a reasonable relationship between tumor purity and allele frequencies of somatic SNVs to accurately estimate the purity. We quantitatively evaluate the performance of STIC on both simulation and real sequencing datasets, the results of which indicate that STIC outperforms competing methods.

Breast cancer is the most common invasive cancer with the highest cancer occurrence in females. Handheld ultrasound is one of the most efficient ways to identify and diagnose the breast cancer. The area and the shape information of a lesion is very helpful for clinicians to make diagnostic decisions. In this study we propose a new deep-learning scheme, semi-pixel-wise cycle generative adversarial net (SPCGAN) for segmenting the lesion in 2D ultrasound. The method takes the advantage of a fully convolutional neural network (FCN) and a generative adversarial net to segment a lesion by using prior knowledge. We compared the proposed method to a fully connected neural network and the level set segmentation method on a test dataset consisting of 32 malignant lesions and 109 benign lesions. Our proposed method achieved a Dice similarity coefficient (DSC) of 0.92 while FCN and the level set achieved 0.90 and 0.79 respectively. Particularly, for malignant lesions, our method increases the DSC (0.90) of the fully connected neural network to 0.93 significantly (p < 0.001). The results show that our SPCGAN can obtain robust segmentation results. The framework of SPCGAN is particularly effective when sufficient training samples are not available compared to FCN. Our proposed method may be used to relieve the radiologists' burden for annotation.

Clustering tumor metastasis samples from gene expression data at the whole genome level remains an arduous challenge, in particular, when the number of experimental samples is small and the number of genes is huge. We focus on the prediction of the epithelial-mesenchymal transition (EMT), which is an underlying mechanism of tumor metastasis, here, rather than tumor metastasis itself, to avoid confounding effects of uncertainties derived from various factors. In this paper, we propose a novel model in predicting EMT based on multidimensional scaling (MDS) strategies and integrating entropy and random matrix detection strategies to determine the optimal reduced number of dimension in low dimensional space. We verified our proposed model with the gene expression data for EMT samples of breast cancer and the experimental results demonstrated the superiority over state-of-the-art clustering methods. Furthermore, we developed a novel feature extraction method for selecting the significant genes and predicting the tumor metastasis. The source code is available at "https://github.com/yushanqiu/yushan.qiu-szu.edu.cn".

Genome Rearrangements are events that affect large stretches of genomes during evolution. Many mathematical models have been used to estimate the evolutionary distance between two genomes based on genome rearrangements. However, most of them focused on the (order of the) genes of a genome, disregarding other important elements in it. Recently, researchers have shown that considering regions between each pair of genes, called intergenic regions, can enhance distance estimation in realistic data. Two of the most studied genome rearrangements are the reversal, which inverts a sequence of genes, and the transposition, which occurs when two adjacent gene sequences swap their positions inside the genome. In this work, we study the transposition distance between two genomes, but we also consider intergenic regions, a problem we name Sorting by Intergenic Transpositions. We show that this problem is NP-hard and propose two approximation algorithms, with factors 3.5 and 2.5, considering two distinct definitions for the problem. We also investigate the signed reversal and transposition distance between two genomes considering their intergenic regions. This second problem is called Sorting by Signed Intergenic Reversals and Intergenic Transpositions. We show that this problem is NP-hard and develop two approximation algorithms, with factors 3 and 2.5. We check how these algorithms behave when assigning weights for genome rearrangements. Finally, we implemented all these algorithms and tested them on real and simulated data.

Knowledge of protein functions plays an important role in biology and medicine. With the rapid development of high-throughput technologies, a huge number of proteins have been discovered. However, there are a great number of proteins without functional annotations. A protein usually has multiple functions and some functions or biological processes require interactions of a plurality of proteins. Additionally, Gene Ontology provides a useful classification for protein functions and contains more than 40,000 terms. We propose a deep learning framework called DeepGOA to predict protein functions with protein sequences and protein-protein interaction (PPI) networks. For protein sequences, we extract two types of information: sequence semantic information and subsequence-based features. We use the word2vec technique to numerically represent protein sequences, and utilize a Bi-directional Long and Short Time Memory (Bi-LSTM) and multi-scale convolutional neural network (multi-scale CNN) to obtain the global and local semantic features of protein sequences, respectively. Additionally, we use the InterPro tool to scan protein sequences for extracting subsequence-based information, such as domains and motifs. Then, the information is plugged into a neural network to generate high-quality features. For the PPI network, the Deepwalk algorithm is applied to generate its embedding information of PPI. Then the two types of features are concatenated together to predict protein functions. To evaluate the performance of DeepGOA, several different evaluation methods and metrics are utilized. The experimental results show that DeepGOA outperforms DeepGO and BLAST.

The past two decades have witnessed significant advances in high-throughput “omics” technologies such as genomics, proteomics, metabolomics, transcriptomics and radiomics. These technologies have enabled simultaneous measurement of the expression levels of tens of thousands of features from individual patient samples and have generated enormous amounts of data that require analysis and interpretation. One specific area of interest has been in studying the relationship between these features and patient outcomes, such as overall and recurrence-free survival, with the goal of developing a predictive “omics” profile. Large-scale studies often suffer from the presence of a large fraction of censored observations and potential time-varying effects of features, and methods for handling them have been lacking. In this paper, we propose supervised methods for feature selection and survival prediction that simultaneously deal with both issues. Our approach utilizes continuum power regression (CPR) - a framework that includes a variety of regression methods - in conjunction with the parametric or semi-parametric accelerated failure time (AFT) model. Both CPR and AFT fall within the linear models framework and, unlike black-box models, the proposed prognostic index has a simple yet useful interpretation. We demonstrate the utility of our methods using simulated and publicly available cancer genomics data.

Glycosylation of proteins in eukaryote cells is an important and complicated post-translation modification due to its pivotal role and association with crucial physiological functions within most of the proteins. Identification of glycosylation sites in a polypeptide chain is not an easy task due to multiple impediments. Analytical identification of these sites is expensive and laborious. There is a dire need to develop a reliable computational method for precise determination of such sites which can help researchers to save time and effort. Herein, we propose a novel predictor namely iGlycoS-PseAAC by integrating the Chou's Pseudo Amino Acid Composition (PseAAC) and relative/absolute position-based features. The self-consistency results show that the accuracy revealed by the model using the benchmark dataset for prediction of O-linked glycosylation having serine sites is 98.8 percent. The overall accuracy of predictor achieved through 10-fold cross validation by combining the positive and negative results is 97.2 percent. The overall accuracy achieved through Jackknife test is 96.195 percent by aggregating of all the prediction results. Thus the proposed predictor can help in predicting the O-linked glycosylated serine sites in an efficient and accurate way. The overall results show that the accuracy of the iGlycoS-PseAAC is higher than the existing tools.

Mammalian Target of Rapamycin (mTOR) is a Ser/Thr protein kinase, and its role is integral to the autophagy pathway in cancer. Targeting mTOR for therapeutic interventions in cancer through autophagy pathway is challenging due to the dual roles of autophagy in tumor progression. The architecture of mTOR reveals two complexes – mTORC1 and mTORC2, each having multiple protein subunits. mTOR kinase inhibitors target the structurally and functionally similar catalytic subunits of both mTORC1 and mTORC2. In this paper, we have explored two different categories of molecular features – descriptors and fingerprints for developing predictive models using machine learning techniques. Random Forest variable importance measures and autoencoders are used to identify molecular descriptors and fingerprints, respectively. We have built various predictive models using identified features and their combination for predicting mTOR kinase inhibitors. Finally, the best model based on the Mathew correlation co-efficient value over the validation dataset is selected for screening kinase SARfari bioactivity dataset. In this study, we have identified twenty best performing descriptors for predicting mTOR kinase inhibitors. To the best of our knowledge, it is the first study on integrating traditional machine learning and deep learning-based approaches for feature extraction to predict mTOR kinase inhibitors.

X-ray crystallography is the most popular approach for analyzing protein 3D structure. However, the success rate of protein crystallization is very low (2-10 percent). To reduce the cost of time and resources, lots of computation-based methods are developed to detect the protein crystallization. Improving the accuracy of predicting protein crystallization is very important for the determination of protein structure by X-ray crystallography. At present, many machine learning methods are used to predict protein crystallization. In this article, we propose a Fuzzy Support Vector Machine based on Linear Neighborhood Representation (FSVM-LNR) to predict the crystallization propensity of proteins. Proteins are represented by three types of features (PsePSSM, PSSM-DWT, MMI-PS), and these features are serially combined and fed into FSVM-LNR. FSVM-LNR can filter outliers by membership score, which is calculated via reconstruction residuals of $k$ k nearest samples. To evaluate the performance of our predictive model, we test FSVM-LNR on the datasets of TRAIN3587, TEST3585 and TEST500. Our method achieves better Mathew’s correlation coefficient (MCC) on TRAIN3587 (MCC: 0.56) and TEST3585 (MCC: 0.58). Although the performance of independent test is not the best on TEST500, FSVM-LNR also has a certain predictability (MCC: 0.70) in the identification of protein crystallization. The good performance on the datasets proves the effectiveness of our method and the better performance on large datasets further demonstrates the stability and superiority of our method.

In the past few years, the prediction models have shown remarkable performance in most biological correlation prediction tasks. These tasks traditionally use a fixed dataset, and the model, once trained, is deployed as is. These models often encounter training issues such as sensitivity to hyperparameter tuning and "catastrophic forgetting" when adding new data. However, with the development of biomedicine and the accumulation of biological data, new predictive models are required to face the challenge of adapting to change. To this end, we propose a computational approach based on Broad learning system (BLS) to predict potential disease-associated miRNAs that retain the ability to distinguish prior training associations when new data need to be adapted. In particular, we are introducing incremental learning to the field of biological association prediction for the first time and proposed a new method for quantifying sequence similarity. In the performance evaluation, the AUC in the 5-fold cross-validation was 0.9400 +/- 0.0041. To better assess the effectiveness of MISSIM, we compared it with various classifiers and former prediction models. Its performance is superior to the previous method. Besides, the case study on identifying miRNAs associated with breast neoplasms, lung neoplasms and esophageal neoplasms show that 34, 36 and 35 out of the top 40 associations predicted by MISSIM are confirmed by recent biomedical resources. These results provide ample convincing evidence of this approach have potential value and prospect in promoting biomedical research productivity.

The paper presents a novel method for processing uncertain data of Phasor measurement unit (PMU) modules first time in the literature using Fuzzy Reasoning Petri net (FPN). It addresses several key issues such as exploitation of Petri net representation from operating state of PMU to its failure state whereas Fuzzy logic is used to deal with the uncertain data of PMU modules. Sprouting tree, an information flow path, of PMU failure is drawn due to various components and estimation accuracy can be enhanced by integration of more truthiness input data. Fault tree diagram, Fuzzy Petri net model (FPN), production rule sets for PMU are developed and finally degree of truthiness of proposition is computed from sprouting tree. Fuzzy logic reasoning is used for routing the sprouting tree whereas Petri net is employed for dynamics of states due to failure of modules of PMU. The fusion of two technologies is made for the dynamic response, processing and reasoning to sprouting tree information flow from operating state to unavailability of PMU. The research work is useful to pinpoint the weakness in design of modules of PMU and to assess its reliability.

Despite fluorescent cell-labelling being widely employed in biomedical studies, some of its drawbacks are inevitable, with unsuitable fluorescent probes or probes inducing a functional change being the main limitations. Consequently, the demand for and development of label-free methodologies to classify cells is strong and its impact on precision medicine is relevant. Towards this end, high-throughput techniques for cell mechanical phenotyping have been proposed to get a multidimensional biophysical characterization of single cells. With this motivation, our goal here is to investigate the extent to which an unsupervised machine learning methodology, which is applied exclusively on morpho-rheological markers obtained by real-time deformability and fluorescence cytometry (RT-FDC), can address the difficult task of providing label-free discrimination of reticulocytes from mature red blood cells. We focused on this problem, since the characterization of reticulocytes (their percentage and cellular features) in the blood is vital in multiple human disease conditions, especially bone-marrow disorders such as anemia and leukemia. Our approach reports promising label-free results in the classification of reticulocytes from mature red blood cells, and it represents a step forward in the development of high-throughput morpho-rheological-based methodologies for the computational categorization of single cells. Besides, our methodology can be an alternative but also a complementary method to integrate with existing cell-labelling techniques., 13 pages, 3 figures, 4 tables

Recent advances in imaging genetics make it possible to combine different types of data including medical images like functional magnetic resonance imaging (fMRI) and genetic data like single nucleotide polymorphisms (SNPs) for comprehensive diagnosis of mental disorders. Understanding complex interactions among these heterogeneous data may give rise to a new perspective, while at the same time demand statistical models for their integration. Various graphical models have been proposed for the study of interaction or association networks with continuous, binary, and count data as well as the mixture of them. However, limited efforts have been made for the multinomial case, for instance, SNP data. Our goal is therefore to fill the void by developing a graphical model for the integration of fMRI image and SNP data, which can provide deeper understanding of the unknown neurogenetic mechanism. In this article, we propose a latent Gaussian copula model for mixed data containing multinomial components. We assume that the discrete variable is obtained by discretizing a latent (unobserved) continuous variable and then create a semi-rank based estimator of the graph structure. The simulation results demonstrate that the proposed latent correlation has more steady and accurate performance than several existing methods in detecting graph structure. When applying to a real schizophrenia data consisting of SNP array and fMRI image collected by the Mind Clinical Imaging Consortium (MCIC), the proposed method reveals a set of distinct SNP-brain associations, which are verified to be biologically significant. The proposed model is statistically promising in handling mixed types of data including multinomial components, which can find widespread applications. To promote reproducible research, the R code is available at https://github.com/Aiying0512/LGCM .

Recently, it was confirmed that ACE2 is the receptor of SARS-CoV-2, the pathogen causing the recent outbreak of severe pneumonia around the world. It is confused that ACE2 is widely expressed across a variety of organs and is expressed moderately but not highly in lung, which, however, is the major infected organ. Therefore, we hypothesized that there could be some other genes playing key roles in the entry of SARS-CoV-2 into human cells. Here we found that AGTR2 (angiotensin II receptor type 2), a G-protein coupled receptor, has interaction with ACE2 and is highly expressed in lung with a high tissue specificity. More importantly, simulation of 3D structure based protein-protein interaction reveals that AGTR2 shows a higher binding affinity with the Spike protein of SARS-CoV-2 than ACE2 (energy: -8.2 vs. -5.1 [kcal/mol]). A number of compounds, biologics and traditional Chinese medicine that could decrease the expression level of AGTR2 were predicted. Finally, we suggest that AGTR2 could be a putative novel gene for the entry of SARS-CoV-2 into human cells, which could provide different insight for the research of SARS-CoV-2 proteins with their receptors.

Accurate and sensitive identification of peptides from MS/MS spectra is a very challenging problem in computational shotgun proteomics. To tackle this problem, spectral library search has been one of the competitive solutions. However, most existing library search tools were developed on the basis of one peptide per spectrum, which prevents them from working properly on chimeric spectra where two or more peptides are co-fragmented. In this work, we present a new library search tool called ChimST, which is particularly capable of reliably identifying multiple peptides from a chimeric spectrum. It starts with associating each query MS/MS spectrum with MS precursor features. For each precursor feature, there is a list of peptide candidates extracted from an input spectral library. Then, it takes one peptide candidate from each associated feature and scores how well they could collectively interpret the query spectrum. The highest-scoring set of peptide candidates are finally reported as the identification of the query spectrum. Our experimental tests show that ChimST could significantly outperform the three state-of-the-art library search tools, SpectraST, reSpect, and MSPLIT, in terms of the numbers of both peptide-spectrum matches and unique peptides, especially when the acquisition isolation window is broad.

Drug–target interaction prediction is important for drug development and drug repurposing. Many computational methods have been proposed for drug–target interaction prediction due to their potential to the time and cost reduction. In this review, we introduce the molecular docking and machine learning-based methods, which have been widely applied to drug–target interaction prediction. Particularly, machine learning-based methods are divided into different types according to the data processing form and task type. For each type of method, we provide a specific description and propose some solutions to improve its capability. The knowledge of heterogeneous network and learning to rank are also summarized in this review. As far as we know, this is the first comprehensive review that summarizes the knowledge of heterogeneous network and learning to rank in the drug–target interaction prediction. Moreover, we propose three aspects that can be explored in depth for future research.

Sample size calculation for adequate power analysis is critical in optimizing RNA-seq experimental design. However, the complexity increases for directly estimating sample size when taking into consideration confounding covariates. Although a number of approaches for sample size calculation have been proposed for RNA-seq data, most ignore any potential heterogeneity. In this study, we implemented a simulation-based and confounder-adjusted method to provide sample size recommendations for RNA-seq differential expression analysis. The data was generated using Monte Carlo simulation, given an underlined distribution of confounding covariates and parameters for a negative binomial distribution. The relationship between the sample size with the power and parameters, such as dispersion, fold change and mean read counts, can be visualized. We demonstrate that the adjusted sample size for a desired power and type one error rate of α is usually larger when taking confounding covariates into account. More importantly, our simulation study reveals that sample size may be underestimated by existing methods if a confounding covariate exists in RNA-seq data. Consequently, this underestimate could affect the detection power for the differential expression analysis. Therefore, we introduce confounding covariates for sample size estimation for heterogeneous RNA-seq data.

Background Identifying differentially expressed genes between the same or different species is an urgent demand for biological and medical research. For RNA-seq data, systematic technical effects and different sequencing depths are usually encountered when conducting experiments. Normalization is regarded as an essential step in the discovery of biologically important changes in expression. The present methods usually involve normalization of the data with a scaling factor, followed by detection of significant genes. However, more than one scaling factor may exist because of the complexity of real data. Consequently, methods that normalize data by a single scaling factor may deliver suboptimal performance or may not even work.The development of modern machine learning techniques has provided a new perspective regarding discrimination between differentially expressed (DE) and non-DE genes. However, in reality, the non-DE genes comprise only a small set and may contain housekeeping genes (in same species) or conserved orthologous genes (in different species). Therefore, the process of detecting DE genes can be formulated as a one-class classification problem, where only non-DE genes are observed, while DE genes are completely absent from the training data. Results In this study, we transform the problem to an outlier detection problem by treating DE genes as outliers, and we propose a scaling-free minimum enclosing ball (SFMEB) method to construct a smallest possible ball to contain the known non-DE genes in a feature space. The genes outside the minimum enclosing ball can then be naturally considered to be DE genes. Compared with the existing methods, the proposed SFMEB method does not require data normalization, which is particularly attractive when the RNA-seq data include more than one scaling factor. Furthermore, the SFMEB method could be easily extended to different species without normalization. Conclusions Simulation studies demonstrate that the SFMEB method works well in a wide range of settings, especially when the data are heterogeneous or biological replicates. Analysis of the real data also supports the conclusion that the SFMEB method outperforms other existing competitors. The R package of the proposed method is available at https://bioconductor.org/packages/MEB.

Background Taxonomic assignment is a key step in the identification of human viral pathogens. Current tools for taxonomic assignment from sequencing reads based on alignment or alignment-free k-mer approaches may not perform optimally in cases where the sequences diverge significantly from the reference sequences. Furthermore, many tools may not incorporate the genomic coverage of assigned reads as part of overall likelihood of a correct taxonomic assignment for a sample. Results In this paper, we describe the development of a pipeline that incorporates a multi-task learning model based on convolutional neural network (MT-CNN) and a Bayesian ranking approach to identify and rank the most likely human virus from sequence reads. For taxonomic assignment of reads, the MT-CNN model outperformed Kraken 2, Centrifuge, and Bowtie 2 on reads generated from simulated divergent HIV-1 genomes and was more sensitive in identifying SARS as the closest relation in four RNA sequencing datasets for SARS-CoV-2 virus. For genomic region assignment of assigned reads, the MT-CNN model performed competitively compared with Bowtie 2 and the region assignments were used for estimation of genomic coverage that was incorporated into a naïve Bayesian network together with the proportion of taxonomic assignments to rank the likelihood of candidate human viruses from sequence data. Conclusions We have developed a pipeline that combines a novel MT-CNN model that is able to identify viruses with divergent sequences together with assignment of the genomic region, with a Bayesian approach to ranking of taxonomic assignments by taking into account both the number of assigned reads and genomic coverage. The pipeline is available at GitHub via https://github.com/MaHaoran627/CNN_Virus.

Background Drug repositioning, meanings finding new uses for existing drugs, which can accelerate the processing of new drugs research and development. Various computational methods have been presented to predict novel drug–disease associations for drug repositioning based on similarity measures among drugs and diseases. However, there are some known associations between drugs and diseases that previous studies not utilized. Methods In this work, we develop a deep gated recurrent units model to predict potential drug–disease interactions using comprehensive similarity measures and Gaussian interaction profile kernel. More specifically, the similarity measure is used to exploit discriminative feature for drugs based on their chemical fingerprints. Meanwhile, the Gaussian interactions profile kernel is employed to obtain efficient feature of diseases based on known disease-disease associations. Then, a deep gated recurrent units model is developed to predict potential drug–disease interactions. Results The performance of the proposed model is evaluated on two benchmark datasets under tenfold cross-validation. And to further verify the predictive ability, case studies for predicting new potential indications of drugs were carried out. Conclusion The experimental results proved the proposed model is a useful tool for predicting new indications for drugs or new treatments for diseases, and can accelerate drug repositioning and related drug research and discovery.

Motivation The circadian rhythm drives the oscillatory expression of thousands of genes across all tissues. The recent revolution in high-throughput transcriptomics, coupled with the significant implications of the circadian clock for human health, has sparked an interest in circadian profiling studies to discover genes under circadian control. Result We present TimeCycle: a topology-based rhythm detection method designed to identify cycling transcripts. For a given time-series, the method reconstructs the state space using time-delay embedding, a data transformation technique from dynamical systems theory. In the embedded space, Takens’ theorem proves that the dynamics of a rhythmic signal will exhibit circular patterns. The degree of circularity of the embedding is calculated as a persistence score using persistent homology, an algebraic method for discerning the topological features of data. By comparing the persistence scores to a bootstrapped null distribution, cycling genes are identified. Results in both synthetic and biological data highlight TimeCycle’s ability to identify cycling genes across a range of sampling schemes, number of replicates and missing data. Comparison to competing methods highlights their relative strengths, providing guidance as to the optimal choice of cycling detection method. Availabilityand implementation A fully documented open-source R package implementing TimeCycle is available at: https://nesscoder.github.io/TimeCycle/. Supplementary information Supplementary data are available at Bioinformatics online.

The Valais Blackneck goat is a Swiss goat breed with a characteristic coat color phenotype. Before the revision of the breed standard in 1938, 4 different color varieties of Valais goats were known. Besides Blackneck animals resembling the modern breed standard, the brown and white Copperneck goat, the white Capra Sempione, and the greyish Grüenochte comprised the historic Valais goats. The brown pigmentation of Copperneck goats had previously been traced back to an introgression of a mutant TYRP1 allele from Toggenburg goats. In the present study, we identified additional introgression events of distinct ASIP alleles causing the remaining 2 rare coat color patterns within the Valais Blackneck goat breed. We identified the introgression of the AWt allele from Appenzell or Saanen goats in white Capra Sempione goats. Similarly, introgression of the Apc allele from Peacock goats resulted in the greyish Grüenochte phenotype. These results demonstrate past hybridization events between breeds that are separated today. A perfect genotype-phenotype association in 393 Valais goats supported the causality of the genotyped variants for the different coat color phenotypes. Our study gives insights into the introgression of functionally relevant copy number variant (CNV) alleles controlling pigmentation between goat breeds with strikingly different coat color patterns.

The holistic Unani medicine is fundamentally different from the reductionist conventional medicine. It asserts the self-integration amongst its basic disciplines, without considering them underlying principles of Unani system of medicine cannot be understood. The diagnosis, selection of drugs, and plan of treatment is also overlooked. Unani scholars attribute health to the functions or actions of the body in a normal way. The constitution is considered as the result of need based on the amalgamation of Arkan (primordial essence). Umoor Tabiya (basic principles) interact at many levels and manifest into Kaifiyyat (Mizaj), Akhlat (Humour), Arwaah (Pneuma), Quwa (faculties), and Tabiyat (Physis) which need to be understood properly for effective management and diagnosis of disease in Unani medicine as well as its treatment. Ilmul Asbab is applied in the prevention of disease as well as in disease causation. In Unani medicine, there should be conformity in between, Asbab (causes), Alamaat (symptoms), and therapeutics. Therefore; the treatment strategy needs the knowledge of Ilmul Asbab. This paper will examine the basic relationship amongst disciplines i.e. Basic principles, diagnosis, and principles of treatment. It will attempt to illustrate the need for awareness of the basic principles of health and disease for a physician for effective management of disease which is what Unani medicine claims to be holistic.

Numerous protein sequences simultaneously exist at multiple subcellular localizations and exhibit multiple functions. Multi-function characterization of the Unknown Protein Sequences (UPS) is useful for analyzing multi-symptom diseases and multi-target drugs. In this work, a multisite subcellular localization model is proposed for the multi-function characterization of UPS using augmented features and algorithm adoption multilabel classifiers. Protein sequence features are augmented with physicochemical and evolutionary properties of amino acid residues as feature vectors while preserving the sequence-order-information and protein-residue-properties. Less discriminative and redundant features are discarded from the feature vector using Multilabel Linear Discriminant Analysis (MLDA). Two different multisite datasets, Gram-Positive (ML_G+) and Gram-Negative (ML_G−) are used for the experimental work, where multiple locative protein sequences with single-label are transformed into a unique multilabel protein sequence. Preprocessed feature vectors of ML_G+ and ML_G− are used separately to train multilabel-classifiers such as Decision Tree (ML_C4.5), K-Nearest Neighbor (ML_kNN), Multi Layer Perceptron (MLP), Extra Tree (ExTr) and Random Forest (RF) using fivefold cross-validation. After that validated multisite model has been utilized for the prediction of single as well as multiple functions of the UPS. The model achieved an accuracy of 94.23% for ML_G+ and 82.77% with ML_G− through known protein sequences using MLP, while for UPS accuracy is 77.50% for ML_G+ using MLP and ExTr, and 54.28% for ML_G− through ML_kNN.

Pathogenicity-related studies are of great importance in understanding the pathogenesis of complex diseases and improving the level of clinical medicine. This work proposed a bioinformatics scheme to analyze cancer-related gene mutations, and try to figure out potential genes associated with diseases from the protein domain-domain interaction network. Herein, five measures of the principle of centrality lethality had been adopted to implement potential correlation analysis, and prioritize the significance of genes. This method was further applied to KEGG pathway analysis by taking the malignant melanoma as an example. The experimental results show that 25 domains can be found, and 18 of them have high potential to be pathogenically important related to malignant melanoma. Finally, a web-based tool, named Human Cancer Related Domain Interaction Network Analyzer, is developed for potential pathogenic genes prioritization for 26 types of human cancers, and the analysis results can be visualized and downloaded online.

Human gait recognition has numerous challenges due to view angle changing, human dressing, bag carrying, and pedestrian walking speed, etc. In order to increase gait recognition accuracy under these circumstances, in this paper we propose a method for gait recognition based on a self-adaptive hidden Markov model (SAHMM). First, we present a feature extraction algorithm based on local gait energy image (LGEI) and construct an observation vector set. By using this set, we optimize parameters of the SAHMM-based method for gait recognition. Finally, the proposed method is evaluated extensively based on the CASIA Dataset B for gait recognition under various conditions such as cross view, human dressing, or bag carrying, etc. Furthermore, the generalization ability of this method is verified based on the OU-ISIR Large Population Dataset. Both experimental results show that the proposed method exhibits superior performance in comparison with those existing methods.

Breast tumor classification with multiple medical reports such as B-ultrasound, Mammography (X-ray) and Nuclear Magnetic Resonance Imaging (MRI) is crucial to the intelligent cancer diagnosis system. Unlike the other domain texts, the medical reports have latent hierarchical syntactic structures and have hidden rich semantic information about the entities and relationships, which poses a great challenge of breast cancer classification. In this article, we proposed a Knowledge-powered Deep Breast Tumor Classification model (KDBTC), which takes the semantic information as a kind of prior knowledge and incorporated it into deep neural networks. Specially, our proposed model first uses Hierarchical Attention Bidirectional Recurrent Neural Networks (HA-BiRNNs) to encode the syntax-aware representation of medical reports in a hierarchical way. In the HA-BiRNN, a hierarchical neural network structure, consisting in two encoder layers of BiRNN (Bidirectional Recurrent Neural Networks), mirrors the hierarchical structure of medical reports, and a hierarchical attention mechanism, consisting of two levels attentions, attends to important elements within clinical report with word-level attention and sentence-level attention. Secondly, our model obtains the semantic information relevant to the medical reports from the clinical domain semantic tree, and encodes the semantic representation of medical reports by using Tree Structured Recurrent Neural Network with gated recursive units (Tree-GRUs). Finally, we classify breast tumors by combining both the syntax and semantic representations of medical reports. We evaluate our method on the real-world breast cancer medical reports, and results show that our method achieves higher performance on breast cancer classification.