Discovering combinatorial interactions in survival data

Abstract

Motivation: Although several methods exist to relate high-dimensional gene expression data to various clinical phenotypes, finding combinations of features in such input remains a challenge, particularly when fitting complex statistical models such as those used for survival studies.Results: Our proposed method builds on existing ‘regularization path-following’ techniques to produce regression models that can extract arbitrarily complex patterns of input features (such as gene combinations) from large-scale data that relate to a known clinical outcome. Through the use of the data’s structure and itemset mining techniques, we are able to avoid combinatorial complexity issues typically encountered with such methods, and our algorithm performs in similar orders of duration as single-variable versions. Applied to data from various clinical studies of cancer patient survival time, our method was able to produce a number of promising gene-interaction candidates whose tumour-related roles appear confirmed by literature.Availability: An R implementation of the algorithm described in this article can be found at https://github.com/david-duverle/regularisation-path-followingContact: dave.duverle@aist.go.jpSupplementary information: Supplementary data are available at Bioinformatics online.

Efficient model selection for predictive pattern mining model by safe pattern pruning

“…In this paper, we consider a class of machine learning models called predictive pattern mining. 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 …”

Identifying statistically significant combinatorial markers for survival analysis

“…With this setting, we can test for association between the failure of samples and the grouping of samples. Most existing methods support their results with p -values computed using log-rank, like in the work of duVerle et al [ 11 ]. To find marker combinations, their method treats combinations as covariates and integrates penalized Cox regression analysis with significant pattern mining to find combinatorial interactions.…”

“…By modifying LAMP, the procedure becomes more suitable for survival data, which generally involves censored information, while enabling us to identify high order combinations without dealing with issues raised by test multiplicity. Similar to [ 11 ], our approach sets no limit on the order of the detected interactions. But unlike them, it does not require training of algorithm that causes possible overfitting of data.…”

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

“…Even though these models are primarily used for main effect selections, there is an increasing interest in incorporating interactions [ 25 - 27 ]. When there is no a priori knowledge, such approaches either require the interactions to be formed by variables that represent main effects or that interaction terms are created by combining the covariates in a certain way, e.g., by producing all distinct two-way interactions (or by coarsening the input space before producing the interactions [ 28 ]). The first route can lead to false negatives even if the true interactions have relevant marginal effects, and the second one neglects the fact that it is frequently either not feasible or computationally too expensive to consider all possible interactions.…”