Combining Semantic Similarity and GO Enrichment for Computation of Functional Similarity

AbstractFunctional similarity between genes is widely used in many bioinformatics applications including detecting molecular pathways, finding co-expressed genes, predicting protein-protein interactions, and prioritization of candidate genes. Methods evaluating functional similarity of genes are mostly based on semantic similarity of gene ontology (GO) terms. Though there are hundreds of functional similarity measures available in the literature, none of them considers the enrichment of the GO terms by the querying gene pair. We propose a novel method to incorporate GO enrichment into the existing functional similarity measures. Our experiments show that the inclusion of gene enrichment significantly improves the performance of 44 widely used functional similarity measures, especially in the prediction of sequence homologies, gene expression correlations, and protein-protein interactions.Software availabilityThe software (python code) and all the benchmark datasets evaluation (R script) are available at https://gitlab.com/liuwt/EnrichFunSim.

Download Full-text

GO2Vec: transforming GO terms and proteins to vector representations via graph embeddings

BMC Genomics ◽

10.1186/s12864-019-6272-2 ◽

2019 ◽

Vol 20 (S9) ◽

Cited By ~ 3

Author(s):

Xiaoshi Zhong ◽

Rama Kaalia ◽

Jagath C. Rajapakse

Keyword(s):

Information Content ◽

Semantic Similarity ◽

Protein Interactions ◽

Large Scale ◽

Functional Similarity ◽

Experimental Results ◽

Graph Embeddings ◽

Protein Protein Interactions ◽

Vector Representations ◽

Go Terms

Abstract Background Semantic similarity between Gene Ontology (GO) terms is a fundamental measure for many bioinformatics applications, such as determining functional similarity between genes or proteins. Most previous research exploited information content to estimate the semantic similarity between GO terms; recently some research exploited word embeddings to learn vector representations for GO terms from a large-scale corpus. In this paper, we proposed a novel method, named GO2Vec, that exploits graph embeddings to learn vector representations for GO terms from GO graph. GO2Vec combines the information from both GO graph and GO annotations, and its learned vectors can be applied to a variety of bioinformatics applications, such as calculating functional similarity between proteins and predicting protein-protein interactions. Results We conducted two kinds of experiments to evaluate the quality of GO2Vec: (1) functional similarity between proteins on the Collaborative Evaluation of GO-based Semantic Similarity Measures (CESSM) dataset and (2) prediction of protein-protein interactions on the Yeast and Human datasets from the STRING database. Experimental results demonstrate the effectiveness of GO2Vec over the information content-based measures and the word embedding-based measures. Conclusion Our experimental results demonstrate the effectiveness of using graph embeddings to learn vector representations from undirected GO and GOA graphs. Our results also demonstrate that GO annotations provide useful information for computing the similarity between GO terms and between proteins.

Download Full-text

Discovering novel protein–protein interactions by measuring the protein semantic similarity from the biomedical literature

Journal of Bioinformatics and Computational Biology ◽

10.1142/s0219720014420086 ◽

2014 ◽

Vol 12 (06) ◽

pp. 1442008 ◽

Cited By ~ 4

Author(s):

Jung-Hsien Chiang ◽

Jiun-Huang Ju

Keyword(s):

Semantic Similarity ◽

Protein Interactions ◽

Similarity Measures ◽

Biomedical Literature ◽

Biological Research ◽

Protein Protein Interactions ◽

Automated Identification ◽

Learning Classifier ◽

Novel Method ◽

Novel Protein

Protein–protein interactions (PPIs) are involved in the majority of biological processes. Identification of PPIs is therefore one of the key aims of biological research. Although there are many databases of PPIs, many other unidentified PPIs could be buried in the biomedical literature. Therefore, automated identification of PPIs from biomedical literature repositories could be used to discover otherwise hidden interactions. Search engines, such as Google, have been successfully applied to measure the relatedness among words. Inspired by such approaches, we propose a novel method to identify PPIs through semantic similarity measures among protein mentions. We define six semantic similarity measures as features based on the page counts retrieved from the MEDLINE database. A machine learning classifier, Random Forest, is trained using the above features. The proposed approach achieve an averaged micro-F of 71.28% and an averaged macro-F of 64.03% over five PPI corpora, an improvement over the results of using only the conventional co-occurrence feature (averaged micro-F of 68.79% and an averaged macro-F of 60.49%). A relation-word reinforcement further improves the averaged micro-F to 71.3% and averaged macro-F to 65.12%. Comparing the results of the current work with other studies on the AIMed corpus (ranging from 77.58% to 85.1% in micro-F, 62.18% to 76.27% in macro-F), we show that the proposed approach achieves micro-F of 81.88% and macro-F of 64.01% without the use of sophisticated feature extraction. Finally, we manually examine the newly discovered PPI pairs based on a literature review, and the results suggest that our approach could extract novel protein–protein interactions.

Download Full-text

Selection of GO-Based Semantic Similarity Measures through AMDE for Predicting Protein-Protein Interactions

Swarm, Evolutionary, and Memetic Computing - Lecture Notes in Computer Science ◽

10.1007/978-3-642-27242-4_7 ◽

2011 ◽

pp. 55-62

Author(s):

Anirban Mukhopadhyay ◽

Moumita De ◽

Ujjwal Maulik

Keyword(s):

Semantic Similarity ◽

Protein Interactions ◽

Similarity Measures ◽

Protein Protein Interactions ◽

Selection Of

Download Full-text

Gene Functional Similarity Analysis by Definition-based Semantic Similarity Measurement of GO Terms

Advances in Artificial Intelligence - Lecture Notes in Computer Science ◽

10.1007/978-3-319-06483-3_18 ◽

2014 ◽

pp. 203-214 ◽

Cited By ~ 4

Author(s):

Ahmad Pesaranghader ◽

Ali Pesaranghader ◽

Azadeh Rezaei ◽

Danoosh Davoodi

Keyword(s):

Semantic Similarity ◽

Functional Similarity ◽

Similarity Measurement ◽

Similarity Analysis ◽

Semantic Similarity Measurement ◽

Go Terms

Download Full-text

Identification of the Disrupted Pathways Associated with Periodontitis Based on Human Pathway Network

Infection International ◽

10.1515/ii-2017-0143 ◽

2016 ◽

Vol 5 (4) ◽

pp. 93-98

Author(s):

Wen Sun ◽

Lin Han ◽

Wenmao Xu ◽

Yazhen Sun

Keyword(s):

Protein Interactions ◽

Threshold Value ◽

Normal Subjects ◽

Protein Protein Interactions ◽

Pathway Network ◽

Differential Expression Genes ◽

Weight Value ◽

Limma Package ◽

Novel Method ◽

Cognitively Normal Subjects

AbstractObjective: The objective of this work is to search for a novel method to explore the disrupted pathways associated with periodontitis (PD) based on the network level.Methods: Firstly, the differential expression genes (DEGs) between PD patients and cognitively normal subjects were inferred based on LIMMA package. Then, the protein-protein interactions (PPI) in each pathway were explored by Empirical Bayesian (EB) co-expression program. Specifically, we determined the 100th weight value as the threshold value of the disrupted pathways of PPI by constructing the randomly model and confirmed the weight value of each pathway. Meanwhile, we dissected the disrupted pathways under the weight value > the threshold value. Pathways enrichment analyses of DEGs were carried out based on Expression Analysis Systematic Explored (EASE) test. Finally, the better method was selected based on the more rich and significant obtained pathways by comparing the two methods.Results: After the calculation of LIMMA package, we estimated 524 DEGs in all. Then we determined 0.115222 as the threshold value of the disrupted pathways of PPI. When the weight value>0.115222, there were 258 disrupted pathways of PPI enriched in. Additionally, we observed those 524 DEGs that were enriched in 4 pathways under EASE=0.1.Conclusion: We proposed a novel network method inferring the disrupted pathway for PD. The disrupted pathways might be underlying biomarkers for treatment associated with PD.

Download Full-text