Metagenomics and metatranscriptomics: windows on CF-associated viral and microbial communities

Abstract

Background: Samples collected from CF patient airways often contain large amounts of host-derived nucleic acids that interfere with recovery and purification of microbial and viral nucleic acids. This study describes metagenomic and metatranscriptomic methods that address these issues.

Methods: Microbial and viral metagenomes, and microbial metatranscriptomes, were successfully prepared from sputum samples from five adult CF patients.

Results: Contaminating host DNA was dramatically reduced in the metagenomes. Each CF patient presented a unique microbiome; in some Pseudomonas aeruginosa was replaced by other opportunistic bacteria. Even though the taxonomic composition of the microbiomes is very different, the metabolic potentials encoded by the community are very similar. The viral communities were dominated by phages that infect major CF pathogens. The metatranscriptomes reveal differential expression of encoded metabolic potential with changing health status.

Conclusions: Microbial and viral metagenomics combined with microbial transcriptomics characterize the dynamic polymicrobial communities found in CF airways, revealing both the taxa present and their current metabolic activities. These approaches can facilitate the development of individualized treatment plans and novel therapeutic approaches.

FIGURE 1

Workflow for the preparation of CF sputum samples for microbiome, virome, and metatranscriptome sequencing.

FIGURE 2

Taxonomic analysis of CF viromes. (a) Putative host range profiles for phage communities. Each bar represents the sum of the normalized abundance values for all phage genotypes with the same putative bacterial host. Only the top 21 hosts are shown. (b) Nucleotide-level alignment of CF1-E virome sequences against a region of the Streptococcus phage Dp-1 genome. Depth of coverage was based on 90% nucleotide identity. The colors represent different nucleotides and demonstrate alignment quality. (c) Coverage plot of Acinetobacter sp. SUN resistance plasmid pRAY recovered from CF1-E and CF5-A.

FIGURE 3

Taxonomic analysis of the microbial communities in three CF patients across multiple time points. (a) Species-level comparisons between microbiomes. Identification was based on unique best hits using BLASTn against the NCBI nucleotide database. All species shown from the same genus are assigned similar colors. (b) Sequence coverage of the Rothia mucilaginosa DY-18 genome by reads from the CF1-E microbiome. (c) Genus-level comparisons between microbiomes and metatranscriptomes. (The CF4-A taxonomy is not shown here because an rRNA removal kit was used during metatranscriptome preparation.)

FIGURE 4

Evaluation of the effects of nebulization and rRNA depletion on the relative amounts of rRNA and non-rRNA in metatranscriptomes. (a) Effects of nebulization. All samples were treated by the Ambion rRNA depletion kits. (b) Comparison of rRNA depletion methods. “Ambion” method uses a combination of MICROBEnrich™ + MICROBExpress™; “Epicentre” method uses Ribo-Zero™ rRNA Removal kit (Epidemiology version).

FIGURE 5

Comparison of KEGG metabolic pathways identified in viromes, microbiomes, and metatranscriptomes as shown by multidimensional scaling (MDS). Grouping by Partitioning Around Medoids (PAM) clustering placed all samples in the appropriate cluster with the exception of the CF4-C metatranscriptome. CF1-A was omitted from both analyses due to insufficient data.