Prediction and validation of association between microRNAs and diseases by multipath methods

Zeng, Xiangxiang; Zhang, Xuan; Liao, Yuanlu; Pan, Linqiang

doi:10.1016/j.bbagen.2016.03.016

Cited by 43 publications

(24 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Zeng et al 30. proposed two multipath methods to predict disease related genes based on gene-disease heterogeneous network and then applied to predict miRNA-disease associations31, and achieved good results. Unfortunately, such machine-learning-based approaches face a common limitation; the negative training samples consisting of non-association between miRNAs and diseases do not demonstrate sufficient statistical confidence because an association was not observed in a biological experiment that cannot draw a conclusion indicating no association between them.…”

mentioning

confidence: 99%

Network Consistency Projection for Human miRNA-Disease Associations Inference

Liao

et al. 2016

Sci Rep

View full text Add to dashboard Cite

Prediction and confirmation of the presence of disease-related miRNAs is beneficial to understand disease mechanisms at the miRNA level. However, the use of experimental verification to identify disease-related miRNAs is expensive and time-consuming. Effective computational approaches used to predict miRNA-disease associations are highly specific. In this study, we develop the Network Consistency Projection for miRNA-Disease Associations (NCPMDA) method to reveal the potential associations between miRNAs and diseases. NCPMDA is a non-parametric universal network-based method that can simultaneously predict miRNA-disease associations in all diseases but does not require negative samples. NCPMDA can also confirm the presence of miRNAs in isolated diseases (diseases without any known miRNA association). Leave-one-out cross validation and case studies have shown that the predictive performance of NCPMDA is superior over that of previous method.

show abstract

mentioning

confidence: 99%

Network Consistency Projection for Human miRNA-Disease Associations Inference

Liao

et al. 2016

Sci Rep

View full text Add to dashboard Cite

show abstract

“…To further confirm the value of our approach, we also used the mean percentile ranking (MPR), an evaluation index based on recall, to evaluate the performance of the algorithm. This evaluation index has been applied in recommendation algorithm and analyses of the performance for predicting drug-targets (Hu et al, 2008;Johnson, 2014;Li et al, 2015;Ding et al, 2017;Hao et al, 2019;Liu et al, 2019b;Liu et al, 2019c; and disease biomarkers (Chen et al, 2016;Zeng et al, 2016;Hong et al, 2019;Xu et al, 2019). For each disease, the genes were ranked in descending order according to the calculated gene-disease predictive value.…”

Section: Evaluation Indexes and Methodsmentioning

confidence: 99%

Pathogenic Gene Prediction Algorithm Based on Heterogeneous Information Fusion

et al. 2020

View full text Add to dashboard Cite

Complex diseases seriously affect people's physical and mental health. The discovery of disease-causing genes has become a target of research. With the emergence of bioinformatics and the rapid development of biotechnology, to overcome the inherent difficulties of the long experimental period and high cost of traditional biomedical methods, researchers have proposed many gene prioritization algorithms that use a large amount of biological data to mine pathogenic genes. However, because the currently known genedisease association matrix is still very sparse and lacks evidence that genes and diseases are unrelated, there are limits to the predictive performance of gene prioritization algorithms. Based on the hypothesis that functionally related gene mutations may lead to similar disease phenotypes, this paper proposes a PU induction matrix completion algorithm based on heterogeneous information fusion (PUIMCHIF) to predict candidate genes involved in the pathogenicity of human diseases. On the one hand, PUIMCHIF uses different compact feature learning methods to extract features of genes and diseases from multiple data sources, making up for the lack of sparse data. On the other hand, based on the prior knowledge that most of the unknown gene-disease associations are unrelated, we use the PU-Learning strategy to treat the unknown unlabeled data as negative examples for biased learning. The experimental results of the PUIMCHIF algorithm regarding the three indexes of precision, recall, and mean percentile ranking (MPR) were significantly better than those of other algorithms. In the top 100 global prediction analysis of multiple genes and multiple diseases, the probability of recovering true gene associations using PUIMCHIF reached 50% and the MPR value was 10.94%. The PUIMCHIF algorithm has higher priority than those from other methods, such as IMC and CATAPULT.

show abstract

“…Machine learning methods have also entered the eld of bioinformatics research. [40][41][42] Support vector machines (SVMs) were used by Jiang et al, 43 Xu et al, 44 Zeng et al 45 and Wang et al, 46 a logistic model tree was used by Wang et al, 47 and a decision tree was used by Zhao et al; 48 these are excellent classication tools with global optimality and better generalization abilities to predict potential disease-related candidate miRNAs, but such methods require known negative sample information related to disease-related miRNAs that is difficult to obtain. In order to solve the problem of negative sample acquisition, Chen et al 49 used a regularized least squares approach to optimize similarity networks of miRNAs and diseases, respectively, and the nal miRNA-disease associations were linear weightings of miRNA similarity scores and disease similarity scores.…”

Section: Introductionmentioning

confidence: 99%