Gene-Gene Interactions Detection Using a Two-stage Model

Jerry Wang and Jae Hoon Sul, two lab alumni, published a paper introducing a new a two-stage model software for detecting associations between traits and pairs of SNPs using a threshold-based efficient pairwise association approach (TEPAA).  The method is significantly faster than the traditional approach of performing an association test with all pairs of SNPs.  In the first stage, the method performs the single marker test on all individual SNPs and selects a subset of SNPs that exceed a certain SNP-specific predetermined significance threshold for further consideration. In the second stage, individual SNPs that are selected in the first stage are paired with each other, and we perform the pairwise association test on those pairs.
The key insight of the approach is that the joint distribution is derived between the association statistics of single SNP and the association statistics of pairs of SNPs. This joint distribution provides guarantees that the statistical power of our approach will closely approximate the brute force approach. Then you can accurately compute the analytical power of our two-stage model and compare it to the power of the brute force approach. (See the Figure) Hence, the method chooses as few SNPs as possible in the first stage while achieving almost the same power as the brute force approach.
The power loss region of the threshold-based efficient pairwise association approach (TEPAA). The contour lines represent the probability density function of the multivariate normal distribution (MVN).  T1(subscript) is the threshold for the first stage.  Any SNP with a higher significance than T1 will be passed on to the second stage.  T2(subscript) is the threshold for significance of the pairwise test.  The area surrounded by the red rectangle corresponds to the power loss region.

The power loss region of the threshold-based efficient pairwise association approach (TEPAA). The contour lines represent the probability density function of the multivariate normal distribution (MVN). T1(subscript) is the threshold for the first stage. Any SNP with a higher significance than T1 will be passed on to the second stage. T2(subscript) is the threshold for significance of the pairwise test. The area surrounded by the red rectangle corresponds to the power loss region.

Jerry and Jae Hoon demonstrate the utility of TEPAA applied to the Northern Finland Birth Cohort (Rantakallio, 1969; Jarvelin et al., 2004).  From their analysis, they observe that the thresholds that control the power loss of the two-stage approach depend on the minor allele frequency (MAF) of the SNPs. In particular, more common SNPs can be filtered out with less significant thresholds than rare SNPs. In order to efficiently implement TEPAA using MAF dependent thresholds for each pair, we group the SNPs into bins based on their MAFs to apply the correct thresholds to each possible pair. After disregarding rare variants with MAF <  0.05, they categorize all common SNPs into nine bins according to their MAF, with step size 0.05. Each pair of SNPs would have two thresholds, one for each SNP in the first stage.  We precompute the first-stage thresholds for each combination of two MAFs in order to achieve 1% power loss,while achieving high cost savings. We sort the SNPs within each bin by their association statistics and use binary search to rapidly obtain the set of SNPs above a single threshold to efficiently implement the first stage of our method.

Read our full paper here:

Sorry, no publications matched your criteria.

Writing Tips: Methods Overview

What are the interesting computational ideas underlying a new computational method?  What are the intuitions behind the method?  How is the method related to other methods?  These are the key question that papers which describe new computational methods should be answering.
Unfortunately, most papers describing new computational methods don’t explicitly address these questions due to constraints of the journal styles.  Introduction of methods papers often have a only few sentences about the method.  The Methods section typically has many more details but has very little discussion of the underlying ideas.   Understanding what is interesting about a method is left completely to the readers imagination.  Often, the journals request that the Results section precede the Methods section which then makes understanding the results very difficult without the reader reading the sections of the paper out of order.  Authors can appeal to the journal to have the Methods section first, but this is also not a good solution since there are many details in the Methods such as descriptions of the datasets which take away from the flow of the paper.
In order to avoid these problems, in our papers, we make the first subsection of the Results section of the paper a “Methods Overview.”  In this section, we describe the method in terms of the high level ideas and typically include as a figure a small example which we utilize the help the reader understand the example.   The goal of this section is to give enough details that the readers can then follow the rest of the Results section without requiring looking at the Methods section.  A well written Methods Overview will make it much easier for the reader to follow the actual Methods section.
These sections and examples are designed to be self contained and should be in a language appropriate for a general audience.  In fact, some of the blog posts are almost verbatim copies of the Methods Overview sections of some of our recent papers.  For example, see these blog posts on GRAT and Genome Reassembly.
Another way to think of what to put in the Methods Overview section is what you would explain in a talk about the method.  Often presentations on computational methods have excellent slides showing intuitions and very clear examples.  The place to put that kind of material is in the Methods Overview.  Remember, in your paper you must give a compelling argument as to WHY your method is interesting. If your readers don’t understand the intuitions underlying your work, they will never appreciate it.
I’m sure you may be asking, “Isn’t this a little redundant?” What I’m proposing here may be a bit repetitive, with a methods overview section and a methods section later in the paper.  But they serve different purposes.  With a well written Methods Overview section, a reader can stop after the Results section and understand most of your paper.  The Methods section then only becomes important for someone who wants to understand all of the details.