Skip to main content
Fig. 4 | Genome Medicine

Fig. 4

From: The lung microbiome, peripheral gene expression, and recurrence-free survival after resection of stage II non-small cell lung cancer

Fig. 4

Peripheral blood gene expression related to survival in stage II NSCLC patients. a For genes selected >125 times in 500 × 10-fold cross-validated elastic-net penalized Cox regression and with FDR-adjusted q < 0.20, we show number of times selected out of 500 times, and the hazard ratio (95% CI) from Cox proportional hazards regression of RFS, DFS, or OS on log2-transformed gene expression, adjusted for age, sex, race, histology, smoking status, and chemotherapy. b Gene ontology (GO) enrichment for survival-related genes. GO terms from the biological process or molecular function ontologies with p < 0.001 from Fisher’s exact test are shown. c, d Partial Spearman’s correlations between relative abundance of survival-related taxa and functional pathways in c normal lung or d tumor and survival-related gene expression. Correlations were adjusted for age, sex, race, histology, smoking status, and chemotherapy. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. e Area under the time-dependent ROC curve (AUC) for the four models at different follow-up times. Model 1 included standard covariates only (age, sex, race, histology, smoking status, and chemotherapy); model 2 (microbiome model) included standard covariates and clr-transformed Clostridiales and Bacteroidales abundance in the normal lung; model 3 (gene model) included standard covariates and log2-transformed peripheral expression of IFITM2, TAP1, TAPBP, and CSF2RB; and model 4 (microbiome and gene model) included standard covariates plus the microbiome and gene variables. Solid lines (circles) represent the time-dependent AUCs from “timeROC” R package; shaded area represents the 95% confidence intervals determined from 1000 bootstraps

Back to article page