Methodology
Spatially Uniform ReliefF (SURF) for computationally-efficient filtering of gene-gene interactions
-
* Corresponding author: Jason H Moore jason.h.moore@dartmouth.edu
1 Department of Genetics, Norris Cotton Cancer Center, Dartmouth Medical School, Lebanon, NH, USA
2 Department of Community and Family Medicine, Dartmouth Medical School, Lebanon, NH, USA
3 Department of Computer Science, University of New Hampshire, Lebanon, NH, USA
4 Department of Computer Science, University of Vermont, Burlington, VT, USA
5 Translational Genomics Research Institute, Phoenix, AZ, USA
BioData Mining 2009, 2:5 doi:10.1186/1756-0381-2-5
Published: 22 September 2009Additional files
Additional file 1:
Appendix. This is an appendix to accompany the manuscript that includes additional theoretical analysis of the Relief algorithms discussed in the manuscript.
Format: PDF Size: 137KB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 2:
Epistasis models. These are the epistasis models used in our data simulation.
Format: PDF Size: 47KB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 3:
Significance of differences with a sample size of 800. This is a plot showing the significance of statistical results for the situation where there are 400 cases and 400 control individuals. These plots follow the example shown in Figure 2. Pairwise comparisons are made between each pair of methods at the 99th, 95th, and 75th percentiles. ReliefF, SURF, TuRF, and SURF & TuRF are labeled R, S, T, and ST respectively. Significance is illustrated with levels of grey (i.e. light grey indicates 0.01 < p ≤ 0.05, dark grey indicates 0.001 < p ≤ 0.01, and black indicates p ≤ 0.001).
Format: PDF Size: 27KB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 4:
Significance of differences with a sample size of 1600. This is a plot showing the significance of statistical results for the situation where there are 800 cases and 800 control individuals. These plots follow the example shown in Figure 2. Pairwise comparisons are made between each pair of methods at the 99th, 95th, and 75th percentiles. ReliefF, SURF, TuRF, and SURF&TuRF are labeled R, S, T, and ST respectively. Significance is illustrated with levels of grey (i.e. light grey indicates 0.01 < p ≤ 0.05, dark grey indicates 0.001 < p ≤ 0.01, and black indicates p ≤ 0.001).
Format: PDF Size: 27KB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 5:
Significance of differences with a sample size of 3200. This is a plot showing the significance of statistical results for the situation where there are 1600 cases and 1600 control individuals. These plots follow the example shown in Figure 2. Pairwise comparisons are made between each pair of methods at the 99th, 95th, and 75th percentiles. ReliefF, SURF, TuRF, and SURF&TuRF are labeled R, S, T, and ST respectively. Significance is illustrated with levels of grey (i.e. light grey indicates 0.01 < p ≤ 0.05, dark grey indicates 0.001 < p ≤ 0.01, and black indicates p ≤ 0.001).
Format: PDF Size: 27KB Download file
This file can be viewed with: Adobe Acrobat Reader