Format

Send to

Choose Destination
Int J Mol Sci. 2014 Sep 29;15(10):17601-21. doi: 10.3390/ijms151017601.

Hope for GWAS: relevant risk genes uncovered from GWAS statistical noise.

Author information

1
Instituto Nacional de Saúde Doutor Ricardo Jorge, Av. Padre Cruz, Lisboa 1649-016, Portugal. ccorreia@igc.gulbenkian.pt.
2
Instituto Gulbenkian de Ciência, Oeiras 2780-156, Portugal. ydiekmann@igc.gulbenkian.pt.
3
Instituto Nacional de Saúde Doutor Ricardo Jorge, Av. Padre Cruz, Lisboa 1649-016, Portugal. astrid.vicente@insa.min-saude.pt.
4
Instituto Gulbenkian de Ciência, Oeiras 2780-156, Portugal. jleal@igc.gulbenkian.pt.

Abstract

Hundreds of genetic variants have been associated to common diseases through genome-wide association studies (GWAS), yet there are limits to current approaches in detecting true small effect risk variants against a background of false positive findings. Here we addressed the missing heritability problem, aiming to test whether there are indeed risk variants within GWAS statistical noise and to develop a systematic strategy to retrieve these hidden variants. Employing an integrative approach, which combines protein-protein interactions with association data from GWAS for 6 common diseases, we found that associated-genes at less stringent significance levels (p < 0.1) with any of these diseases are functionally connected beyond noise expectation. This functional coherence was used to identify disease-relevant subnetworks, which were shown to be enriched in known genes, outperforming the selection of top GWAS genes. As a proof of principle, we applied this approach to breast cancer, supporting well-known breast cancer genes, while pinpointing novel susceptibility genes for experimental validation. This study reinforces the idea that GWAS are under-analyzed and that missing heritability is rather hidden. It extends the use of protein networks to reveal this missing heritability, thus leveraging the large investment in GWAS that produced so far little tangible gain.

PMID:
25268625
PMCID:
PMC4227180
DOI:
10.3390/ijms151017601
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Multidisciplinary Digital Publishing Institute (MDPI) Icon for PubMed Central
Loading ...
Support Center