Investigating the impact human protein–protein interaction networks have on disease-gene analysis

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Advances in high-throughput technologies along with the curation of small-scale experiments has aided in the construction of reference maps of the interactome. These maps are critical to our understanding of genotype-phenotype relationships and disease. However, our knowledge of disease associated genes and the map of the human interactome still remains incomplete. In this study we investigate whether protein–protein interaction networks (PPINs) constructed from either experimental or curated data have an impact upon disease network analysis. An integrative network-driven framework is implemented to integrate diverse heterogeneous data including: gene-expression, PPIN, ontology-based similarity, degree connectivity and betweenness centrality measures to uncover potential Alzhemier disease (AD) candidate genes. Two PPINs have been selected and constructed from (1) experimental high-throughput data and (2) literature-curated sources. Only a marginal overlap of protein pairs between the two PPINs (305 protein pairs) was observed. A total of 17 significant AD gene candidate genes were identified using the literature derived PPIN compared to 20 genes using the PPIN constructed from high-throughput data. Both approaches correctly identified the AD susceptible TRAF1, a critical regulator of cerebral ischaemia–reperfusion injury and neuronal death. Biological process enrichment analysis revealed genes candidates from the literature based PPIN are modulated in AD pathogenesis such as neuron differentiation and involved in KEGG pathways such as neurotrophin signaling pathways. Tissue specific analysis revealed 48 % of AD gene candidates obtained from the literature curated PPIN were expressed in tissues where AD is observed compared to 19 % of gene candidates extracted using the high-throughput PPIN.
LanguageEnglish
Pages455-464
JournalInternational Journal of Machine Learning and Cybernetics
Volume9
Issue number3
Early online date13 Feb 2016
DOIs
Publication statusPublished - Mar 2018

Fingerprint

Genes
Proteins
Throughput
Tissue
Electric network analysis
Gene expression
Neurons
Ontology

Keywords

  • network analysis
  • data integration
  • protein interaction networks
  • Alzheimer's Disease

Cite this

@article{64c39ba771174543bc60feb0917c8cbe,
title = "Investigating the impact human protein–protein interaction networks have on disease-gene analysis",
abstract = "Advances in high-throughput technologies along with the curation of small-scale experiments has aided in the construction of reference maps of the interactome. These maps are critical to our understanding of genotype-phenotype relationships and disease. However, our knowledge of disease associated genes and the map of the human interactome still remains incomplete. In this study we investigate whether protein–protein interaction networks (PPINs) constructed from either experimental or curated data have an impact upon disease network analysis. An integrative network-driven framework is implemented to integrate diverse heterogeneous data including: gene-expression, PPIN, ontology-based similarity, degree connectivity and betweenness centrality measures to uncover potential Alzhemier disease (AD) candidate genes. Two PPINs have been selected and constructed from (1) experimental high-throughput data and (2) literature-curated sources. Only a marginal overlap of protein pairs between the two PPINs (305 protein pairs) was observed. A total of 17 significant AD gene candidate genes were identified using the literature derived PPIN compared to 20 genes using the PPIN constructed from high-throughput data. Both approaches correctly identified the AD susceptible TRAF1, a critical regulator of cerebral ischaemia–reperfusion injury and neuronal death. Biological process enrichment analysis revealed genes candidates from the literature based PPIN are modulated in AD pathogenesis such as neuron differentiation and involved in KEGG pathways such as neurotrophin signaling pathways. Tissue specific analysis revealed 48 {\%} of AD gene candidates obtained from the literature curated PPIN were expressed in tissues where AD is observed compared to 19 {\%} of gene candidates extracted using the high-throughput PPIN.",
keywords = "network analysis, data integration, protein interaction networks, Alzheimer's Disease",
author = "Fiona Browne and Wang, {Haiying / HY} and Huiru Zheng",
year = "2018",
month = "3",
doi = "10.1007/s13042-016-0503-5",
language = "English",
volume = "9",
pages = "455--464",
journal = "International Journal of Machine Learning and Cybernetics",
issn = "1868-8071",
number = "3",

}

TY - JOUR

T1 - Investigating the impact human protein–protein interaction networks have on disease-gene analysis

AU - Browne, Fiona

AU - Wang, Haiying / HY

AU - Zheng, Huiru

PY - 2018/3

Y1 - 2018/3

N2 - Advances in high-throughput technologies along with the curation of small-scale experiments has aided in the construction of reference maps of the interactome. These maps are critical to our understanding of genotype-phenotype relationships and disease. However, our knowledge of disease associated genes and the map of the human interactome still remains incomplete. In this study we investigate whether protein–protein interaction networks (PPINs) constructed from either experimental or curated data have an impact upon disease network analysis. An integrative network-driven framework is implemented to integrate diverse heterogeneous data including: gene-expression, PPIN, ontology-based similarity, degree connectivity and betweenness centrality measures to uncover potential Alzhemier disease (AD) candidate genes. Two PPINs have been selected and constructed from (1) experimental high-throughput data and (2) literature-curated sources. Only a marginal overlap of protein pairs between the two PPINs (305 protein pairs) was observed. A total of 17 significant AD gene candidate genes were identified using the literature derived PPIN compared to 20 genes using the PPIN constructed from high-throughput data. Both approaches correctly identified the AD susceptible TRAF1, a critical regulator of cerebral ischaemia–reperfusion injury and neuronal death. Biological process enrichment analysis revealed genes candidates from the literature based PPIN are modulated in AD pathogenesis such as neuron differentiation and involved in KEGG pathways such as neurotrophin signaling pathways. Tissue specific analysis revealed 48 % of AD gene candidates obtained from the literature curated PPIN were expressed in tissues where AD is observed compared to 19 % of gene candidates extracted using the high-throughput PPIN.

AB - Advances in high-throughput technologies along with the curation of small-scale experiments has aided in the construction of reference maps of the interactome. These maps are critical to our understanding of genotype-phenotype relationships and disease. However, our knowledge of disease associated genes and the map of the human interactome still remains incomplete. In this study we investigate whether protein–protein interaction networks (PPINs) constructed from either experimental or curated data have an impact upon disease network analysis. An integrative network-driven framework is implemented to integrate diverse heterogeneous data including: gene-expression, PPIN, ontology-based similarity, degree connectivity and betweenness centrality measures to uncover potential Alzhemier disease (AD) candidate genes. Two PPINs have been selected and constructed from (1) experimental high-throughput data and (2) literature-curated sources. Only a marginal overlap of protein pairs between the two PPINs (305 protein pairs) was observed. A total of 17 significant AD gene candidate genes were identified using the literature derived PPIN compared to 20 genes using the PPIN constructed from high-throughput data. Both approaches correctly identified the AD susceptible TRAF1, a critical regulator of cerebral ischaemia–reperfusion injury and neuronal death. Biological process enrichment analysis revealed genes candidates from the literature based PPIN are modulated in AD pathogenesis such as neuron differentiation and involved in KEGG pathways such as neurotrophin signaling pathways. Tissue specific analysis revealed 48 % of AD gene candidates obtained from the literature curated PPIN were expressed in tissues where AD is observed compared to 19 % of gene candidates extracted using the high-throughput PPIN.

KW - network analysis

KW - data integration

KW - protein interaction networks

KW - Alzheimer's Disease

U2 - 10.1007/s13042-016-0503-5

DO - 10.1007/s13042-016-0503-5

M3 - Article

VL - 9

SP - 455

EP - 464

JO - International Journal of Machine Learning and Cybernetics

T2 - International Journal of Machine Learning and Cybernetics

JF - International Journal of Machine Learning and Cybernetics

SN - 1868-8071

IS - 3

ER -