High-dimensional omics data analysis using a variable screening protocol with prior knowledge integration (SKI)

Abstract

Background: High-throughput technology could generate thousands to millions biomarker measurements in one experiment. However, results from high throughput analysis are often barely reproducible due to small sample size. Different statistical methods have been proposed to tackle this "small n and large p" scenario, for example different datasets could be pooled or integrated together to provide an effective way to improve reproducibility. However, the raw data is either unavailable or hard to integrate due to different experimental conditions, thus there is an emerging need to develop a method for "knowledge integration" in high-throughput data analysis.

Results: In this study, we proposed an integrative prescreening approach, SKI, for high-throughput data analysis. A new rank is generated based on two initial ranks: (1) knowledge based rank; and (2) marginal correlation based rank. Our simulation shows the SKI outperforms other methods without knowledge-integration in terms of higher true positive rate given the same number of variables selected. We also applied our method in a drug response study and found its performance to be better than regular screening methods.

Conclusion: The proposed method provides an effective way to integrate knowledge for high-throughput analysis. It could easily implemented with our provided R package named SKI.

Keywords: Dimension reduction; Knowledge integration; SKI; Sure independence screening; Variable selection.

Fig. 1

A brief description of (i)SKI procedure. For each variable, two ranks are generated, one based on prior knowledge (R ₀), the other based on marginal correlation (R ₁). A predefined α, (or estimated based on the dev. ratio) is used to control the weight of prior knowledge. Variables are then sorted by weighted geometric mean of two ranks. SKI first reduces the variable number from p to d, and then a more sophisticated method such as SCAD is used to further refine the model to size d ’ and estimate the parameters. iSKI updates the marginal correlation based rank (R ₁) by regressing residues over the rest p − d ’ variables. The procedure is repeated until the desired number of parameters obtained

Fig. 2

Boxplot of squared error for selumtinib response prediction using two methods. Whiskers indicate min/max, upper box limit 75% percentile, low box limit 25% percentile and line the median