Figure 4.
Identification of differentially abundant NOGs at the level of DNA and transcription. (a) Overlap of NOGs detected with ⩾1 read in ⩾1 sample in metagenomic and metatranscriptomic analyses. (b) Distributions of reads per million (RPM) for NOGs detected in DNA-seq, RNA-seq or both data sets (average across 16 samples). *Wilcoxon rank-sum test P<0.05. (c) Correlation between NOG abundance estimates (log2) of 14 211 NOGs that were detected at an abundance of >0.1 RPM in both data sets. (d) PCA of metagenomeSeq normalised NOG abundances (14 211 NOGs detected in both data sets at >0.1 RPM) in the metagenomic data set. (e) PCA of metagenomeSeq normalised NOG abundances (14 211 NOGs detected in both data sets at >0.1 RPM) in the metatranscriptomic data set. (f) Overlap of NOGs called as differentially abundant (adjusted P<0.05 in H. h+aIL10R (colitis) versus steady state) in metagenomic and metatranscriptomic data sets. (g) Fold enrichment of NOG functional categories in the set of NOGs that were found to be more highly abundant and transcribed in colitis versus steady state in both metagenomic and metatranscriptomic data sets. (h) Normalised RNA abundance of NOGs in the inorganic ion transport and metabolism eggNOG functional category that were found to be more highly abundant and transcribed in colitis versus steady state in both metagenomic and metatranscriptomic data sets. (i) The proportion of metatranscriptomic reads from each genus that contributed to transcription of each NOG identified in (h). Only those genera that contributed >1% of metatranscriptomic reads to any given NOG are shown. (j) Correlation of NOG abundance estimates (log2) across replicate metatranscriptomic data sets. The mean abundance across samples is plotted in each case. (k) Overlap of NOGs called as significantly differentially abundant (Benjamini-Hochberg adjusted P-value <0.05) across replicate metatranscriptomic data sets. The significance of the overlap was calculated using the hypergeometric test.