Fig. 4

Detection of the relative key fungi with NSCLC. (A) Two methods are used to find key differential fungi between two groups. The volcano plot demonstrates the differential abundance of fungi between NSCLC and Non-NSCLC. Points are colored according to the number of log2 fold change with passed thresholds (p < 0.05 and |log2 fold change|> 1) which are calculated by DEseq2. The bar chart of key fungi between two groups which is determined by Random forest model (p < 0.05). (B) The fungi associated with NSCLC progress. We sorted each fungus by the adjusted R² calculated from OLS model. Only the fungi with p < 0.05 and FDR < 0.2 are showed. (C) A venn graph shows the comprehensive results of the three methods. The hypergeometric test is used to check if the five key fungi detected by two methods (DEseq2 and Random Forest) enrich with the NSCLC development. (D) A line chart shows the variation between the log₁₀ mean relative abundance of the A. arborescens and cancer stage. (E) Another cohort is used to confirm the enrichment of A. arborescens in the tissue of NSCLC (Paired Wilcoxon test, p = 0.00005)