Combining techniques for screening and evaluating interaction terms on high-dimensional time-to-event data

Abstract

BackgroundMolecular data, e.g. arising from microarray technology, is often used for predicting survival probabilities of patients. For multivariate risk prediction models on such high-dimensional data, there are established techniques that combine parameter estimation and variable selection. One big challenge is to incorporate interactions into such prediction models. In this feasibility study, we present building blocks for evaluating and incorporating interactions terms in high-dimensional time-to-event settings, especially for settings in which it is computationally too expensive to check all possible interactions.ResultsWe use a boosting technique for estimation of effects and the following building blocks for pre-selecting interactions: (1) resampling, (2) random forests and (3) orthogonalization as a data pre-processing step. In a simulation study, the strategy that uses all building blocks is able to detect true main effects and interactions with high sensitivity in different kinds of scenarios. The main challenge are interactions composed of variables that do not represent main effects, but our findings are also promising in this regard. Results on real world data illustrate that effect sizes of interactions frequently may not be large enough to improve prediction performance, even though the interactions are potentially of biological relevance.ConclusionScreening interactions through random forests is feasible and useful, when one is interested in finding relevant two-way interactions. The other building blocks also contribute considerably to an enhanced pre-selection of interactions. We determined the limits of interaction detection in terms of necessary effect sizes. Our study emphasizes the importance of making full use of existing methods in addition to establishing new ones.

Do little interactions get lost in dark random forests?

“…However, interaction effects might again be masked by marginal effects in that approach. A related idea is to uncover marginal effects in a first step and project the remaining effects on a space orthogonal to the marginal effects, to detect interactions in a second step [ 32 , 33 ]. On a different route, the detection of interactions might be facilitated by developing new pairwise importance measures based on standard random forests [ 34 ].…”

An Update on Statistical Boosting in Biomedicine

“…Only recently has stability selection been adapted to become a tool for variable selection within the boosting framework (e.g., [ 47 ]). Other work in this context is the analysis of the connections between boosting and penalized regression [ 10 ] and the work by Sariyar et al [ 85 ] exploring a combination of boosting and random forest methods.…”

“…The general motivation is similar to the step length modification factor proposed by Sariyar et al [ 57 ]. In another approach, Sariyar et al [ 85 ] combined a likelihood-based boosting approach for the Cox model with random forests in order to screen for interaction effects in high-dimensional data. Hieke et al [ 86 ] combined likelihood-based boosting with resampling to identify prognostic SNPs in potentially small clinical cohorts.…”

Identifying gene‐gene interactions using penalized tensor regression

“…( b ) The second method is the Cox model with modified adaptive Lasso penalization (HierLasso) [11]. ( c ) A boosting interaction screening model based on random forests and the Cox model [29] is adopted and realized using R package sprinter . This method does not account for the hierarchical structure.…”