EMT network-based feature selection improves prognosis prediction in lung adenocarcinoma

Abstract

Various feature selection algorithms have been proposed to identify cancer prognostic biomarkers. In recent years, however, their reproducibility is criticized. The performance of feature selection algorithms is shown to be affected by the datasets, underlying networks and evaluation metrics. One of the causes is the curse of dimensionality, which makes it hard to select the features that generalize well on independent data. Even the integration of biological networks does not mitigate this issue because the networks are large and many of their components are not relevant for the phenotype of interest. With the availability of multi-omics data, integrative approaches are being developed to build more robust predictive models. In this scenario, the higher data dimensions create greater challenges.We proposed a phenotype relevant network-based feature selection (PRNFS) framework and demonstrated its advantages in lung cancer prognosis prediction. We constructed cancer prognosis relevant networks based on epithelial mesenchymal transition (EMT) and integrated them with different types of omics data for feature selection. With less than 2.5% of the total dimensionality, we obtained EMT prognostic signatures that achieved remarkable prediction performance (average AUC values >0.8), very significant sample stratifications, and meaningful biological interpretations. In addition to finding EMT signatures from different omics data levels, we combined these single-omics signatures into multi-omics signatures, which improved sample stratifications significantly. Both single-and multi-omics EMT signatures were tested on independent multi-omics lung cancer datasets and significant sample stratifications were obtained.

Clinical Impact of the Epithelial-Mesenchymal Transition in Lung Cancer as a Biomarker Assisting in Therapeutic Decisions

“…In SCCs, vimentin expression failed to establish an independent prognosis, but high S100A expression and lack of intercellular E-cadherin allowed to predict patients at a high risk of recurrence and poor prognosis [Zhang et al, 2013]. Both in ACs and SCCs, reduced membranous staining of E-cadherin and expression of vimentin were shown to be independent predictors of mortality [Che et al, 2015;Aruga et al, 2018;Shao et al, 2019].…”

“…Using NSCLC cell lines and tumor samples, some studies established EMT signatures. From these NSCLC reports, some markers such as DSP, TJP1, CLDN4, ERBB3, GALNT3, and CDS1 appear commonly associated with CDH1 epithelial-related genes while MMP-2, AXL, ZEB1/2, NRP-1, and TWIST clustered with VIM and FN1 mesenchymal-related genes [Wushou et al, 2014;Zheng and Kang, 2014;Antony and Huang, 2017;Goossens et al, 2017;Wong et al, 2018;Karacosta et al, 2019;Shao et al, 2019].…”