Zhicheng Liu

The analysis of the drug–targets based on the topological properties in the human protein–protein interaction network

Analyzing topological properties of drug-target proteins in the biology network is very helpful in understanding the mechanism of drug action. However, comprehensive studies to elaborately characterize the biological network features of drug-target proteins are still lacking. In this paper, we compared the topological properties of drug–targets with those of the non–drug-target sets, by mapping the drug–targets in DrugBank to the human protein interaction network. The results indicate that the topological properties of drug-targets are significantly distinguishable from those of non–drug-targets. Moreover, the potential possibility of drug-target prediction based on these properties is discussed. All proteins in the interaction network were ranked by their topological properties. Among the top 200 proteins, 94 overlapped with drug-targets in DrugBank and some novel predictions were found to be drug–targets in public literatures and other databases. In conclusion, our method explores the topological properties of drug-targets in the human protein interaction network by exploiting the large–scale drug-targets and protein interaction data.

Topological properties of the drug targets regulated by microRNA in human protein-protein interaction network.

The investigation of topological properties of proteins in protein–protein interaction network (PPIN) has great potentials to identify basic protein functions and mechanisms of action. Based on human PPIN, previous study has shown that the topological properties of drug targets are significantly distinguished from those of proteins that are not targeted by drugs (non-drug-targets). MicroRNAs (miRNAs) are known to regulate gene expression at the post-transcriptional level. To determine whether the differences in topological properties between drug targets and non-drug-targets are dominated by the proteins that are regulated by miRNA, we divided the drug targets into two sets: those are regulated by miRNA (mir-drug-targets) and those are not regulated by miRNA (non-mir-drug-targets). We compared the probability of interactions and five topological properties among the three types of proteins in human PPIN. Our results demonstrated that mir-drug-targets preferentially interact with other mir-drug-targets and tend to be hub-bottlenecks. However, there was no bias on topological properties between non-mir-drug-targets and non-drug-targets. The same topological features are observed among non-drug targets. These findings indicate that miRNA regulation has an important role in human PPIN, and may be useful in the development of novel drugs.

Identifying drug-target proteins based on network features

Proteins rarely function in isolation inside and outside cells, but operate as part of a highly interconnected cellular network called the interaction network. Therefore, the analysis of the properties of drug-target proteins in the biological network is especially helpful for understanding the mechanism of drug action in terms of informatics. At present, no detailed characterization and description of the topological features of drug-target proteins have been available in the human protein-protein interaction network. In this work, by mapping the drug-targets in DrugBank onto the interaction network of human proteins, five topological indices of drug-targets were analyzed and compared with those of the whole protein interactome set and the non-drug-target set. The experimental results showed that drug-target proteins have higher connectivity and quicker communication with each other in the PPI network. Based on these features, all proteins in the interaction network were ranked. The results showed that, of the top 100 proteins, 48 are covered by DrugBank; of the remaining 52 proteins, 9 are drug-target proteins covered by the TTD, Matador and other databases, while others have been demonstrated to be drug-target proteins in the literature.

Determination of the material parameters of four-fibre family model based on uniaxial extension data of arterial walls

The four-fibre family constitutive relation has been used to capture the mechanical behaviour of arterial walls under biaxial loading conditions. This study shows that the material parameters of the four-fibre family model can be determined by uniaxial extension data from the arterial walls. Stochastic optimisation methods were used to determine the material parameters based on uniaxial extension data of the strip samples with circumferential and axial orientations from thoracic aortas and pulmonary arteries of two fresh donation bodies. Moreover, we implemented numerical experiments, in which stress–strain data generated according to different constitutive parameters were treated as mechanical experiment data and went through the same methods as mechanical test data to determine the constitutive parameters. The estimate–effect ratio, defined by the number of data with the precision of estimation less than 0.5% over whole size of data, was applied to demonstrate the feasibility of our method. The material parameters for Chinese thoracic aorta and pulmonary artery were given with , and the minimal estimate–effect ratio in numerical simulations was 97.77%. In conclusion, the four-fibre family model of arterial walls can be determined from uniaxial extension data. Moreover, the four-fibre family six-parameter constitutive model is the best fit to the data from Chinese pulmonary arteries, and the four-fibre family eight-parameter constitutive model is the best fit to the data from Chinese thoracic aortas.

Mining Functional Gene Modules Linked with Rheumatoid Arthritis Using a SNP-SNP Network

The identification of functional gene modules that are derived from integration of information from different types of networks is a powerful strategy for interpreting the etiology of complex diseases such as rheumatoid arthritis (RA). Genetic variants are known to increase the risk of developing RA. Here, a novel method, the construction of a genetic network, was used to mine functional gene modules linked with RA. A polymorphism interaction analysis (PIA) algorithm was used to obtain cooperating single nucleotide polymorphisms (SNPs) that contribute to RA disease. The acquired SNP pairs were used to construct a SNP-SNP network. Sub-networks defined by hub SNPs were then extracted and turned into gene modules by mapping SNPs to genes using dbSNP database. We performed Gene Ontology (GO) analysis on each gene module, and some GO terms enriched in the gene modules can be used to investigate clustered gene function for better understanding RA pathogenesis. This method was applied to the Genetic Analysis Workshop 15 (GAW 15) RA dataset. The results show that genes involved in functional gene modules, such as CD160 (rs744877) and RUNX1 (rs2051179), are especially relevant to RA, which is supported by previous reports. Furthermore, the 43 SNPs involved in the identified gene modules were found to be the best classifiers when used as variables for sample classification.

The correlation of gene expression and co-regulated gene patterns in characteristic KEGG pathways

There is great interest in chromosome- and pathway-based techniques for genomics data analysis in the current work in order to understand the mechanism of disease. However, there are few studies addressing the abilities of machine learning methods in incorporating pathway information for analyzing microarray data. In this paper, we identified the characteristic pathways by combining the classification error rates of out-of-bag (OOB) in random forests with pathways information. At each characteristic pathway, the correlation of gene expression was studied and the co-regulated gene patterns in different biological conditions were mined by Mining Attribute Profile (MAP) algorithm. The discovered co-regulated gene patterns were clustered by the average-linkage hierarchical clustering technique. The results showed that the expression of genes at the same characteristic pathway were approximate. Furthermore, two characteristic pathways were discovered to present co-regulated gene patterns in which one contained 108 patterns and the other contained one pattern. The results of cluster analysis showed that the smallest similarity coefficient of clusters was more than 0.623, which indicated that the co-regulated patterns in different biological conditions were more approximate at the same characteristic pathway. The methods discussed in this paper can provide additional insight into the study of microarray data.

Mining susceptibility gene modules and disease risk genes from SNP data by combining network topological properties with support vector regression

Genome-wide association study is a powerful approach to identify disease risk loci. However, the molecular regulatory mechanisms for most complex diseases are still not well understood. Therefore, further investigating the interplay between genetic factors and biological networks is important for elucidating the molecular mechanisms of complex diseases. Here, we proposed a novel framework to identify susceptibility gene modules and disease risk genes by combining network topological properties with support vector regression from single nucleotide polymorphism (SNP) level. We assigned risk SNPs to genes using the University of California at Santa Cruz (UCSC) genome database, and then mapped these genes to protein-protein interaction (PPI) networks. The gene modules implicated by hub genes were extracted using the PPI networks and the topological property was analyzed for these gene modules. For each gene module, risk feature genes were determined by topological property analysis and support vector regression. As a result, five shared risk feature genes, CD80, EGFR, FN1, GSK3B and TRAF6 were found and proven to be associated with rheumatoid arthritis by previous reports. Our approach showed a good performance in comparison with other approaches and can be used for prioritizing candidate genes associated with complex diseases.

Power type strain energy function model and prediction of the anisotropic mechanical properties of skin using uniaxial extension data.

Many successful models to describe the biomechanical characteristics of planar biological soft tissues are based on strain energy function. However, the parameters in these models are determined by biaxial extension test, which might be difficult to exercise for certain types of soft tissue. This study presents a new constitutive model, the power type strain energy density function model (PTM), and a method to identify its material parameters for rabbit skin using uniaxial extension test of 4-direction strip samples. The abdominal skins from eight rabbits were taken to perform uniaxial tension tests in 7 different directions. The material parameters were identified for each subject based on any 4 out of 7 directions by applying some definite conditions of this issue. For each rabbit, the 35 groups of material parameters were consistent. The 7 material parameters in PTM were identified with root mean square errors <0.061. The results indicate that the material parameters of rabbit skin can be identified from uniaxial extension test data.

Screening for cancer associated MiRNAs through co-gene, co-function and co-pathway analysis

MicroRNAs (miRNAs) though present themselves as a group of non-coding small RNAs play critical roles in many biological and pathological processes. Among which the regulation of human cancer is one of the most excited potentiality. The goal of this study is to obtain miRNAs robustly associated with cancer by screening all of the possible miRNAs/cancer pairs in three consecutive steps. First, in co-gene analysis, gene set enrichment analysis is carried out for all miRNA/cancer pairs. Second, in co-function analysis, information theoretic similarity on GO is calculated for miRNA/cancer pairs screened from the former step. Third, in co-pathway analysis, pathway enrichment analysis is performed for miRNA/cancer pairs screened from the second step. In this study, we totally included 776 miRNAs and 25 cancer types. As a result, 94 miRNAs were identified with robust association with 17 types of cancer. Meanwhile, 83 pathways with relevance to both miRNAs and cancer were also singled out. This framework provides an effective way to narrow down miRNAs for cancer and to pinpoint corresponding pathways.

Three-dimensional Hessian matrix-based quantitative vascular imaging of rat iris with optical-resolution photoacoustic microscopy in vivo

Abstract. For the diagnosis and evaluation of ophthalmic diseases, imaging and quantitative characterization of vasculature in the iris are very important. The recently developed photoacoustic imaging, which is ultrasensitive in imaging endogenous hemoglobin molecules, provides a highly efficient label-free method for imaging blood vasculature in the iris. However, the development of advanced vascular quantification algorithms is still needed to enable accurate characterization of the underlying vasculature. We have developed a vascular information quantification algorithm by adopting a three-dimensional (3-D) Hessian matrix and applied for processing iris vasculature images obtained with a custom-built optical-resolution photoacoustic imaging system (OR-PAM). For the first time, we demonstrate in vivo 3-D vascular structures of a rat iris with a the label-free imaging method and also accurately extract quantitative vascular information, such as vessel diameter, vascular density, and vascular tortuosity. Our results indicate that the developed algorithm is capable of quantifying the vasculature in the 3-D photoacoustic images of the iris in-vivo, thus enhancing the diagnostic capability of the OR-PAM system for vascular-related ophthalmic diseases in vivo.