Metabolome permafrost

This is a repository containing the code for replicating the analysis and the figures for the paper:

Multi-Omics Unveils Feature-Level Microbe-Metabolite Linkages Driving Permafrost Greenhouse Gas Dynamics

Assembly Microbiome

1. R scripts from 1 to 2 were used to calculate the ecological processes contributing to microbiome assembly. The input of these scripts include a phylogenetic tree derived from metagenome assembled geneomes and their abundances. Author Hannah Holland Moritz

Assembly Metabolome

1. R scripts from 1 to 5 were adapted from Danczak et al. 2020, https://github.com/danczakre/Meta-Metabolome_Ecology to calculate the assembly processes contributing to the peat metabolome. The input of the first script is a report of the relative abundances of masses detected through FT-ICR MS. The outputs of each script serve as input of the following ones.

Assembly at feature level: microbiome and metabolome

1. R scripts in these sections were adapted from Danczak et al. 2022, https://github.com/danczakre/betaNTI-feature to calculate the βNTI_feature value within each habitat across the permafrost gradient. This analysis were performed for the microbiome and metabolome separately. The input for the microbiome analysis include a phylogenetic tree derived from metagenome assembled geneomes and their abundances. The input for the metabolome analysis include the dedrogram (TWCD) and processed report derived from the Assembly Metabolome section.

βNTI_feature clustering

1. Cluster_bnti_features.Rmd: Script for clustering metabolites based on their βNTI_feature value and identifying the representative features per each cluster. Inputs are βNTI_feature matrices filtered by habitat. Outputs are list of representative features per habitat (rep_features_{habitat}.csv files) and the βNTI_feature values of the representative features (rep_features_{habitat}_bnti.csv files).

2. Cluster_correlations.Rmd: Script to calculate the Spearman correlations between the consensus βNTI_feature value of each cluster in each sample against microbial βNTI_feature values, microbial abundances, environmental data (i.e., depth, soil temperature, %C %N, and precipitation) and greenhouse gas (CO₂ and CH₄) data. Inputs are the the lists of representative features per habitat generated by the Cluster_bnti_features.Rmd and csv files of the previously mentioned datasets. Outputs are csv files with the results of the correlations with every data type (*_correlations.csv files).

Figures

1. Figure_1.Rmd: Script to generate the complete Figure 1 of the paper, including the comparison of the average within-habitat &beta nearest taxon index (βNTI) and the ecological processes driving the microbiome and metabolome across the permafrost gradient. Inputs for this figure derive from results of the Assembly Microbiome and Assembly Metabolome folders. Also, it includes correlations of the beta nearest taxon index (βNTI) with environmental factors. Output is Figure 1.

2. Figure_2.Rmd: Script to generate the complete Figure 2 of the paper, including the correlation networks of the βNTI_feature clusters, as well as the biogeochemical indexes, metabolomome assembly contribution, and elemental composition of the representative features. Inputs are the correlation results generated by the Cluster_correlations.Rmd script, and the βNTI_feature values of the representative features generated by the Cluster_bnti_features.Rmd script. Output is Figure 2.

3. Figure_3.Rmd: Script to generate the complete Figure 3 of the paper, including the correlation to greenhouse gases (GHG) and functional potential of bacteria associated with the βNTI_feature clusters. Inputs are abundances, βNTI_feature values, and metagenome annotation tables of microbiome performed with DRAM from the bog as well as GHG concentration data. Output is Figure 3.

4. Figure_4A-C_Supplementary_4.Rmd: Script to generate the first half of Figure 4 of the paper, the microbial-metabolite co-occurrence networks within each habitat. Inputs are the network's edge files produced by MENAP and chemical information (e.g., elemental composition) of the metabolites in the networks (node_table_annotated.csv) derived from previous analysis. Additionally this scripts generates Supplementary Figure 4, which includes information about the networks (e.g., number of nodes and edges). Outputs are panels A through C of Figure 4, Supplementary Figure 4, and a table detailing the bacteria present in the networks (networked_bacteria.csv).

5. Figure_4D-F_Supplementary_5.Rmd: Script to generate the other half of Figure 4 of the paper, the Sankey diagrams of the networked bacteria. Inputs are the taxonomical annotation of the networked bacteria (network_nodes_taxonomy.csv), the phylogenetic tree of the bacteria (provided by Hannah Holland-Moritz) and the Sankey diagram inputs and annotation tables produced by METABOLIC (Metabolic_Sankey_diagram_input_*.txt files). Additionally this scripts generates Supplementary Figure 5, which indicates the number of annotated and expressed CAZymes present in the networked bacteria. Output is panels D through F of Figure 4 and Supplementary Figure 5.

6. Supplementary_1.R: Script to generate Supplementary Figure 1 of the paper which includes the βNTI_feature contribution and the relative percentage of metabolite elemental composition within each habitat. The relative abundance of metabolite elemental compositions and the relative abundance of different metabolite classes in the βNTI_feature devided clusters. Output is Supplementary figure 1.

7. Supplementary_2_Supplementary_3.Rmd: Script to generate Supplementary Figure 2 of the paper which includes the Expression in transcripts per million (TPM) of metagenome assembled genomes (MAGs) that correlated with metabolite-BNTI-feature derived clusters 1 in the bog. Script to generate Supplementary Figure 3 Expression in transcripts per million (TPM) of metagenome assembled genomes (MAGs) that correlated with greenhouse gases in the bog. Output is Supplemenatary Figure 2 and 3.

8. Supplementary_6.Rmd: Script to generate Supplementary Figure 6 of the paper which includes the expression in transcripts per million (TPM) of metagenome assembled genomes (MAGs) that form part of the microbe-metabolite networks in the three habitats. Output is supplementary Figure 6.

9. Supplementary_7.Rmd: Script to generate Supplementary Figure 7 of the paper which represents ordination plot of the metabolome and microbiome. Metabolome UniFrac PCoA was generated using the TWCD dendrogram Output is supplementary Figure 7.

10. Supplementary_8.Rmd: Script to generate Supplementary Figure 8 of the paper which represents a mantel correlogram using Pearson’s correlation coefficient of between-OTU environmental optima and between-OTU phylogenetic distance for environmental variables such as soil temperature, average depth, carbon content, nitrogen content, C:N ratio and precipitation (3-day accumulation). Significant correlations observed as solid squares and non-significant correlations as open squares.. Output is supplementary Figure 8.