Artificial neural networks for modeling gene-gene- and gene-environment-interactions in genetic epidemiology
Description
- The investigation of genetic factors is an important research topic in epidemiology. Most relevant from a public health perspective are complex diseases that are characterised by complex pathways involving gene-gene- and gene-environment-interactions. The identification of such pathways requires sophisticated statistical methods that are still in their infancy. Due to their ability in describing complex association structures, directed graphs may represent a suitable means for modelling complex causal pathways. This project investigated the appropriateness of using neural networks for modelling complex pathways in association studies.
The project investigated the ability of artificial neural networks to model and to identify gene-gene and gene-environment interactions. It was showen that neural networks are very well suited to represent different interactions including discrete as well as continuous variables. In a second step, local and global model fit criteria were investigated. For interpreting the parameters, different approaches were studied like e.g. edge reduction systems and the concept of generalized synaptic weights.
The superiority of neural networks over regression models and the multifactor dimensionality reduction (MDR) could be shown. Multi-layer perceptrons are a very useful statistical tool and should be used in genetic-epidemiological studies if only few is known about the relationship between independent and dependent variables.
Within the project, the R-package neuralnet was implemented and published within the Comprehensive R Archive Network (CRAN).
Funding period
- Begin: August 2006
End: July 2009
Sponsor
- German Research Foundation
Contact
Selected project-related publications
- Günther F, Bammann K, Pigeot I. Using a reduced topology of an artificial neural network to identify biological interaction in genetic epidemiology for two-locus disease models. JP Journal of Biostatistics. 2012;7(2):61-76.
- Günther F, Pigeot I, Bammann K. Artificial neural networks modeling gene-environment interaction. BMC Genetics. 2012;13:37.
https://doi.org/10.1186/1471-2156-13-37 - Günther F, Fritsch S. neuralnet: Training of neural networks. R Journal. 2010;2(1):30-38.
https://journal.r-project.org/archive/2010/RJ-2010-006/index.html - Günther F, Wawro N, Bammann K. Neural networks for modeling gene-gene interactions in association studies. BMC Genetics. 2009;10:87.
https://doi.org/10.1186/1471-2156-10-87 - Günther F, Bammann K, Pigeot I. Kantenreduktion bei künstlichen neuronalen Netzen. In: Foraita R, Ziegler A, Hemmelmann C, editors. Biometrische Aspekte der Genomanalyse IV, Schwerpunkt: Epigenetik. Aachen: Shaker Verlag. 2009. S. 74-81
- Günther F, Wawro N, Bammann K. Neural networks for modelling gene-environment interactions. Workshop "Statistical Methods for the Analysis of Gene-Environment Interactions" der Arbeitsgruppe Genetische Epidemiologie der Deutschen Gesellschaft für Epidemiologie (DGEpi) und des Deutschen Krebsforschungszentrums (DKFZ), 15.-16. März 2010, Heidelberg.
- Günther F, Bammann K, Pigeot I. Kantenreduktion bei künstlichen neuronalen Netzen. 55. Biometrisches Kolloquium, 17.-19. März 2009, Hannover.
- Günther F, Bammann K, Pigeot I. Kantenreduktion bei künstlichen neuronalen Netzen. Workshop der Arbeitsgruppe Populationsgenetik und Genomanalyse der Internationalen Biometrischen Gesellschaft (IBS-DR) und des Arbeitskreises Humangenetik der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (GMDS), 16.-17. Februar 2009, Rauischholzhausen.
- Günther F, Wawro N, Bammann K. Using generalized weights of neural networks to identify gene-gene interactions. EUROEPI 2009, European Congress "Epidemiology for Clinical Medicine and Public Health", 26.-30. August 2009, Warsaw, Poland.
- Bammann K, Günther F. Künstliche neuronale Netze zur Modellierung von Gen-Umwelt-Interaktionen. 3. Jahrestagung der Deutschen Gesellschaft für Epidemiologie (DGEpi), 24.-27. September 2008, Bielefeld.
- Günther F, Bammann K, Wawro N. Neural networks modeling gene-gene-interactions. Lifestat 2008, Gemeinsame Tagung der Deutschen Region (IBS-DR), der Region Österreich/Schweiz (IBS-ROeS) und der polnischen Gruppe (IBS-Polen) der Internationalen Biometrischen Gesellschaft (IBS), 10.-13. März 2008, München.
- Bammann K, Günther F, Wawro N. Identifying complex causal pathways with artificial neural networks: a simulation study. Joint Meeting of the UK Society for Social Medicine (SSM) & the International Epidemiological Association (IEA Europe), 12-14 September 2007, Cork, Ireland. (Abstract published in: The Journal of Epidemiology and Community Health. 2007;61(Suppl.1):A16)
- Günther F, Bammann K, Wawro N. Modellierung von Gen-Gen-Interaktionen mit neuronalen Netzen. Workshop "Genetische Epidemiologie". Treffen der norddeutschen Zentren zu genetischer Epidemiologie Lübeck, Kiel, Bremen, 9. Juli 2007, Lübeck.
- Günther F, Bammann K, Wawro N. Neuronale Netze zur Identifizierung von Gen-Gen-Interaktionen. Kongress "Medizin und Gesellschaft. Prävention und Versorgung innovativ-qualitätsgesichert-sozial". Gemeinsame Jahrestagung der DGEpi, DGSMP, GMDS, DGMS, MDK Bayern, bayer. Landesgesundheitsamt, 17.-21. September 2007, Augsburg.
- Bammann K, Foraita R, Pigeot I, Suling M, Günther F. Modelling gene-gene- and gene-environment-interactions with directed graphs. IEA-EEF European Congress of Epidemiology 2006, Epidemiology and Health Care Practice. 28.Juni - 1.Juli 2006, Utrecht, The Netherlands.
- Günther F, Bammann K, Pigeot I. Kantenreduktion bei künstlichen neuronalen Netzen. 3. Jahrestagung der Deutschen Gesellschaft für Epidemiologie (DGEpi), 24.-27. September 2008, Bielefeld.
- Fritsch S, Günther F. neuralnet: Training of neural networks. R package. (Online); 2012.
http://cran.r-project.org/web/packages/neuralnet/index.html