Jiawei Luo

A graph regularized non-negative matrix factorization method for identifying microRNA-disease associations

Motivation MicroRNAs (miRNAs) play crucial roles in post-transcriptional regulations and various cellular processes. The identification of disease-related miRNAs provides great insights into the underlying pathogenesis of diseases at a system level. However, most existing computational approaches are biased towards known miRNA-disease associations, which is inappropriate for those new diseases or miRNAs without any known association information. Results In this study, we propose a new method with graph regularized non-negative matrix factorization in heterogeneous omics data, called GRNMF, to discover potential associations between miRNAs and diseases, especially for new diseases and miRNAs or those diseases and miRNAs with sparse known associations. First, we integrate the disease semantic information and miRNA functional information to estimate disease similarity and miRNA similarity, respectively. Considering that there is no available interaction observed for new diseases or miRNAs, a preprocessing step is developed to construct the interaction score profiles that will assist in prediction. Next, a graph regularized non-negative matrix factorization framework is utilized to simultaneously identify potential associations for all diseases. The results indicated that our proposed method can effectively prioritize disease-associated miRNAs with higher accuracy compared with other recent approaches. Moreover, case studies also demonstrated the effectiveness of GRNMF to infer unknown miRNA-disease associations for those novel diseases and miRNAs. Availability The code of GRNMF is freely available at https://github.com/XIAO-HN/GRNMF/. Supplementary information Supplementary data are available at Bioinformatics online.

Feature selection in machine learning: A new perspective

Abstract High-dimensional data analysis is a challenge for researchers and engineers in the fields of machine learning and data mining. Feature selection provides an effective way to solve this problem by removing irrelevant and redundant data, which can reduce computation time, improve learning accuracy, and facilitate a better understanding for the learning model or data. In this study, we discuss several frequently-used evaluation measures for feature selection, and then survey supervised, unsupervised, and semi-supervised feature selection methods, which are widely applied in machine learning problems, such as classification and clustering. Lastly, future challenges about feature selection are discussed.

Predicting MicroRNA-Disease Associations Using Network Topological Similarity Based on DeepWalk

Recently, increasing experimental studies have shown that microRNAs (miRNAs) involved in multiple physiological processes are connected with several complex human diseases. Identifying human disease-related miRNAs will be useful in uncovering novel prognostic markers for cancer. Currently, several computational approaches have been developed for miRNA-disease association prediction based on the integration of additional biological information of diseases and miRNAs, such as disease semantic similarity and miRNA functional similarity. However, these methods do not work well when this information is unavailable. In this paper, we present a similarity-based miRNA-disease prediction method that enhances the existing association discovery methods through a topology-based similarity measure. DeepWalk, a deep learning method, is utilized in this paper to calculate similarities within a miRNA-disease association network. It shows superior predictive performance for 22 complex diseases, with area under the ROC curve scores ranging from 0.805 to 0.937 by using five-fold cross-validation. In addition, case studies on breast cancer, lung cancer, and prostatic cancer further justify the use of our method to discover latent miRNA-disease pairs.

Prediction of microRNA–disease associations with a Kronecker kernel matrix dimension reduction model

Identifying the associations between human diseases and microRNAs is key to understanding pathogenicity mechanisms and important for uncovering novel prognostic markers. To date, a series of computational approaches have been developed for the prediction of disease–microRNA associations. However, these methods remain difficult to perform satisfactorily for diseases with a few known associated microRNAs. This study introduces a novel computational model, namely, the Kronecker kernel matrix dimension reduction (KMDR) model, for identifying potential microRNA–disease associations. This model combines microRNA space and disease space in a larger microRNA–disease space by using the Kronecker product or the Kronecker sum. The predictive performance of our proposed approach was evaluated and validated based on known association datasets. The experimental results show that KMDR achieves reliable prediction with an average AUC of 0.8320 for 22 complex diseases, which indeed outperforms other competitive methods. Moreover, case studies on kidney cancer, breast cancer, and esophageal cancer further demonstrate the applicability of our method in the identification of new disease–microRNA pairs. The source code of KMDR is freely available at https://github.com/ghli16/KMDR.

Inferring human miRNA functional similarity based on gene ontology annotations

Even though miRNAs participate in many critical biological processes, the most miRNA functions are ambiguous. To understand the functions of miRNAs, biologists focus on comparing targets of miRNAs, and aim to determine the interactions between targets. However, existing methods may not satisfy the demand accurately because of various limitations. Uncovering the functional associations of miRNAs is also a challenge. Here, we developed MFS_GO to infer miRNA functional similarity using their target genes. We apply MFS_GO to the miRNA families and miRNA clusters and validate our results using miRNA expression, miRNA-disease associations and computational complexity. Compared with the miFRunSim, MFS_GO shows its strong performance. We also conducted our method on breast neoplasms, colorectal neoplasms, and ovarian neoplasms. Experimental results showed that MFS_GO can find novel candidate cancer-related miRNAs with no need for presetting the number of known cancer-related miRNAs.

A cell-core-attachment approach for identifying protein complexes in PPI network

With the increasing of available protein-protein interaction (PPI) data can be applied to computational methods to predict protein complexes. Most methods identify dense regions, which correspond to a cluster in the PPI network, as protein complexes. However, these dense regions could be protein complexes or sets of proteins that have circumstantial close interaction, which could affect the prediction accuracy. In this paper, a new algorithm for predicting protein complexes, applied the framework of the cell-core-attachment, is proposed. The triangular structures (cells) are obtained as the center of complex cores using the edge-clustering coefficient. Then the cells are expanded to protein complex cores. Finally, the attachments are appended to their corresponding cores to form the whole protein complexes. We apply the new approach into several yeast protein interaction data. Results show that the predicted protein complexes can match with the benchmark protein complexes better than the classic algorithms. Moreover, most of these protein complexes detected by our algorithm are biologically significant.

Human disease MiRNA inference by combining target information based on heterogeneous manifolds

The emergence of network medicine has provided great insight into the identification of disease-related molecules, which could help with the development of personalized medicine. However, the state-of-the-art methods could neither simultaneously consider target information and the known miRNA-disease associations nor effectively explore novel gene-disease associations as a by-product during the process of inferring disease-related miRNAs. Computational methods incorporating multiple sources of information offer more opportunities to infer disease-related molecules, including miRNAs and genes in heterogeneous networks at a system level. In this study, we developed a novel algorithm, named inference of Disease-related MiRNAs based on Heterogeneous Manifold (DMHM), to accurately and efficiently identify miRNA-disease associations by integrating multi-omics data. Graph-based regularization was utilized to obtain a smooth function on the data manifold, which constitutes the main principle of DMHM. The novelty of this framework lies in the relatedness between diseases and miRNAs, which are measured via heterogeneous manifolds on heterogeneous networks integrating target information. To demonstrate the effectiveness of DMHM, we conducted comprehensive experiments based on HMDD datasets and compared DMHM with six state-of-the-art methods. Experimental results indicated that DMHM significantly outperformed the other six methods under fivefold cross validation and de novo prediction tests. Case studies have further confirmed the practical usefulness of DMHM.

Predicting microRNA-disease associations using label propagation based on linear neighborhood similarity

Interactions between microRNAs (miRNAs) and diseases can yield important information for uncovering novel prognostic markers. Since experimental determination of disease-miRNA associations is time-consuming and costly, attention has been given to designing efficient and robust computational techniques for identifying undiscovered interactions. In this study, we present a label propagation model with linear neighborhood similarity, called LPLNS, to predict unobserved miRNA-disease associations. Additionally, a preprocessing step is performed to derive new interaction likelihood profiles that will contribute to the prediction since new miRNAs and diseases lack known associations. Our results demonstrate that the LPLNS model based on the known disease-miRNA associations could achieve impressive performance with an AUC of 0.9034. Furthermore, we observed that the LPLNS model based on new interaction likelihood profiles could improve the performance to an AUC of 0.9127. This was better than other comparable methods. In addition, case studies also demonstrated our methods outstanding performance for inferring undiscovered interactions between miRNAs and diseases, especially for novel diseases.

A cell-core-attachment approach for identifying protein complexes in yeast protein-protein interaction network

With the increasing of available protein-protein interaction (PPI) data, many computational methods have been explored to identify protein complexes from PPI networks. Majority of algorithms employ the feature of local neighbors to detect local dense subgraphs which correspond to protein complexes. Those approaches neglect the inherent core-attachment structure of protein complexes, which to an extent affect the protein complexes of prediction accuracy. In this paper, we propose a new algorithm for predicting protein complexes, deriving from the framework of the core-attachment. The proposed method first obtains the triangular structures of the core of protein complexes, name as cells, in which the edge-clustering coefficient is used. And then the cells are expanded to protein complex cores based on the closeness. Finally, the attachments are added to their corresponding cores to form the final protein complexes. The experimental results on two yeast PPI data show our method outperform the existing algorithms in terms of matched protein complexes and biological significance using two benchmark data sets.

Gene Identification Based on Geometrical Representation of DNA Sequence

We consider a new 4D representation of DNA sequence, which has the advantage of not only containing all the information in the DNA sequence but also avoiding the overlapping. Based on this representation, we define a new common frequency coefficient and apply it to gene identification. The identification result of the S.cerevisiae genome illustrates the superior performance of our approach.