NERI: network-medicine based integrative approach for disease gene prioritization by relative importance

Autor(es)

Simoes, Sergio N. ^{[1, 2]} ; Martins, Jr., David C. ^[3] ; Pereira, Carlos A. B. ^[1] ; Hashimoto, Ronaldo F. ^[1] ; Brentani, Helena ^{[4, 5, 6]}

Afiliação do(s) autor(es)

^[1] Univ Sao Paulo, Sch Math & Stat, Sao Paulo, SP - Brazil

^[2] Fed Inst Espirito Santo, Serra, ES - Brazil

^[3] Fed Univ ABC, Ctr Math Computat & Cognit, Santo Andre, SP - Brazil

^[4] Univ Sao Paulo, Sch Med, Dept Psychiat, Sao Paulo, SP - Brazil

^[5] Univ Sao Paulo, Res Ctr Neurodev & Mental Hlth, Sao Paulo, SP - Brazil

^[6] Natl Inst Dev Psychiat Children & Adolescents INC, Seo Paulo, SP - Brazil

• Tipo de documento: Artigo Científico
• Fonte: BMC Bioinformatics; v. 16, n. 19 DEC 16 2015.
• Citações Web of Science: 6

Resumo

Background: Complex diseases are characterized as being polygenic and multifactorial, so this poses a challenge regarding the search for genes related to them. With the advent of high-throughput technologies for genome sequencing, gene expression measurements (transcriptome), and protein-protein interactions, complex diseases have been sistematically investigated. Particularly, Protein-Protein Interaction (PPI) networks have been used to prioritize genes related to complex diseases according to its topological features. However, PPI networks are affected by ascertainment bias, in which more studied proteins tend to have more connections, degrading the results quality. Additionally, methods using only PPI networks can provide only static and non-specific results, since the topologies of these networks are not specific of a given disease. Results: The goal of this work is to develop a methodology that integrates PPI networks with disease specific data sources, such as GWAS and gene expression, to find genes more specific of a given complex disease. After the integration of PPI networks and gene expression data, the resulting network is used to connect genes related to the disease through the shortest paths that have the greatest concordance between their gene expressions. Both case and control expression data are used separately and, at the end, the most altered genes between the two conditions are selected. To evaluate the method, schizophrenia was adopted as case study. Conclusion: Results show that the proposed method successfully retrieves differentially coexpressed genes in two conditions, while avoiding the bias from literature. Moreover we were able to achieve a greater concordance in the selection of important genes from different microarray studies of the same disease and to produce a more specific gene set related to the studied disease. (AU)

• Processo FAPESP: 10/52138-8 - Integração de dados na biologia sistêmica: caracterização de fenômenos biológicos a partir de informações estruturais e funcionais
• Beneficiário: Ronaldo Fumio Hashimoto
• Linha de fomento: Auxílio à Pesquisa - Regular
• Processo FAPESP: 11/50761-2 - Modelos e métodos de e-Science para ciências da vida e agrárias
• Beneficiário: Roberto Marcondes Cesar Junior
• Linha de fomento: Auxílio à Pesquisa - Temático