Gene prioritization in Type 2 Diabetes using domain interactions and network analysis

BMC Genomics

Research Article

Nikhil Tandon¹  Rubina Tabassum²  Sreenivas Chavali²  Amitabh Sharma²  Dwaipayan Bharadwaj² 

[1] Department of Endocrinology, All India Institute of Medical Sciences, New Delhi, India;Functional Genomics Unit, Institute of Genomics and Integrative Biology, CSIR, Delhi, India;

关键词: Gene Ontology;    Disease Gene;    Cluster Coefficient;    Positional Candidate;    Disease Candidate;   

DOI  :  10.1186/1471-2164-11-84

 received in 2009-05-19, accepted in 2010-02-02,  发布年份 2010

来源: Springer

【 摘 要 】

BackgroundIdentification of disease genes for Type 2 Diabetes (T2D) by traditional methods has yielded limited success. Based on our previous observation that T2D may result from disturbed protein-protein interactions affected through disrupting modular domain interactions, here we have designed an approach to rank the candidates in the T2D linked genomic regions as plausible disease genes.ResultsOur approach integrates Weight value (Wv) method followed by prioritization using clustering coefficients derived from domain interaction network. Wv for each candidate is calculated based on the assumption that disease genes might be functionally related, mainly facilitated by interactions among domains of the interacting proteins. The benchmarking using a test dataset comprising of both known T2D genes and non-T2D genes revealed that Wv method had a sensitivity and specificity of 0.74 and 0.96 respectively with 9 fold enrichment. The candidate genes having a Wv > 0.5 were called High Weight Elements (HWEs). Further, we ranked HWEs by using the network property-the clustering coefficient (Ci). Each HWE with a Ci < 0.015 was prioritized as plausible disease candidates (HWEc) as previous studies indicate that disease genes tend to avoid dense clustering (with an average Ci of 0.015). This method further prioritized the identified disease genes with a sensitivity of 0.32 and a specificity of 0.98 and enriched the candidate list by 6.8 fold. Thus, from the dataset of 4052 positional candidates the method ranked 435 to be most likely disease candidates. The gene ontology sharing for the candidates showed higher representation of metabolic and signaling processes. The approach also captured genes with unknown functions which were characterized by network motif analysis.ConclusionsPrioritization of positional candidates is essential for cost-effective and an expedited discovery of disease genes. Here, we demonstrate a novel approach for disease candidate prioritization from numerous loci linked to T2D.

【 授权许可】

Unknown   
© Sharma et al; licensee BioMed Central Ltd. 2010. This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.