False Discovery Rate Control for High Dimensional Networks of Quantile Associations Conditioning on Covariates

Abstract

Summary. Motivated by gene coexpression pattern analysis, we propose a novel sample quantile contingency (SQUAC) statistic to infer quantile associations conditioning on covariates. It features enhanced flexibility in handling variables with both arbitrary distributions and complex association patterns conditioning on covariates. We first derive its asymptotic null distribution, and then develop a multiple-testing procedure based on the SQUAC statistic to test simultaneously the independence between one pair of variables conditioning on covariates for all p(p–1)/2 pairs. Here, p is the length of the outcomes and could exceed the sample size. The testing procedure does not require resampling or perturbation and thus is computationally efficient. We prove by theory and numerical experiments that this testing method asymptotically controls the false discovery rate. It outperforms all alternative methods when the complex association patterns exist. Applied to a gastric cancer data set, this testing method successfully inferred the gene coexpression networks of early and late stage patients. It identified more changes in the networks which are associated with cancer survivals. We extend our method to the case that both the length of the outcomes and the length of covariates exceed the sample size, and show that the asymptotic theory still holds.

SifiNet: A robust and accurate method to identify feature gene sets and annotate cells

“…First, SifiNet is a co-expression network-based method; thus, the quality of the estimated co-expression network is crucial to the following steps. To ensure the high-quality, we used quantile association networks to replace the commonly used correlation networks: quantile association networks have been shown to work better for the count data [46], especially for the transcripts counts. We also explored the rigorous theoretical properties of resulting co-expression networks: the mathematical theorems and proofs were provided in the Supplementary note 1 and 2.…”

Distance Assisted Recursive Testing

“…have been proposed and discussed in other papers such as Liu et al (2013) and Xie and Li (2018). After the testing procedure on layer 1, denote the rejected feature set by…”

A Double-Robust Test for High-Dimensional Gene Coexpression Networks Conditioning on Clinical Information

“…There are also many papers focusing on equivalent characterizations of CI, such as Petersen and Hansen (2021) and Cai et al (2022). A number of other methods have addressed the CI test from different perspectives, including the test for the independence of regression residuals (Belloni et al, 2014;Shah et al, 2020), the conditional randomization test by Candes et al (2018), the conditional permutation test by Berrett et al (2020), the conditional quantile association test (Xie & Li, 2018), and the evaluation of the distance between two conditional density/characteristics functions (Runge, 2018;Su & White, 2007;Wang et al, 2015;Wang & Hong, 2018).…”

“…(2018), the conditional permutation test by Berrett et al. (2020), the conditional quantile association test (Xie & Li, 2018), and the evaluation of the distance between two conditional density/characteristics functions (Runge, 2018; Su & White, 2007; Wang et al., 2015; Wang & Hong, 2018).…”