Construction and Functional Analysis of Human Genetic Interaction Networks with Genome-wide Association Data

(arXiv: 1101.3343, pdf)

Gang Fang, Wen Wang, Vanja Paunic, Benjamen Oatley, Majda Haznada, Michael Steinbach, Brian Van Ness, Chad L. Myers and Vipin Kumar

Correspondence: Gang Fang (gangfang cs umn edu)
Last updated: 01/18/2010

Supplementary files for the submission to ISMB 2011 (Results, with consistent observations, that were not included in the submission due to space limit):

[0] Due to the limit of space, we did not include a figure illustrating the trivial BPM pattern that may result from LD structures in SNP data. This figure motivates the second step of the proposed framework for network contruction, i.e. identifying LD block and summarizing SNP-SNP network to LD block-block network.

[1] In Figure 1 in the submission, we show the results on two of the six datasets, Parkinson and M-Survival due to space limit. The figures for all the six datasets can be found here, which give consistent observations over all the six datasets.

[2] In Table 2 in the submission, we show the number of significant BPMs (with FDR <=0.25) discovered from each of the six datasets, with respect to four measures. Due to space limit, we omited two columns for two other measures about the sizes of a BPM, i.e. S_sum and S_max. This because the FDRs with respect to these two measures are significant on three of the six datasets but not the other three.  The full table can be found here. Note that, the highlight is that on each of the first four measures, there are significant BPMs discovered from all the six datasets.

[3] In Figure 4 in the submission, due to space limit, we illustrate two of the discovered BPMs, i.e. one from Parkinson and one from M-Survival . Here, we show one BPM example for each of the six datasets. Note that, there are multiple significant BPMs discovered from each of the six dataset, and for visual illustration, we take one example from each, while Table 2 shows the number of all the significant BPMs discovered from each of the dataset.