Metagenome-assembled microbial genomes from Parkinson's disease fecal samples

Abstract

The human gut microbiome composition has been linked to Parkinson's disease (PD). However, knowledge of the gut microbiota on the genome level is still limited. Here we performed deep metagenomic sequencing and binning to build metagenome-assembled genomes (MAGs) from 136 human fecal microbiomes (68 PD samples and 68 control samples). We constructed 952 non-redundant high-quality MAGs and compared them between PD and control groups. Among these MAGs, there were 22 different genomes of Collinsella and Prevotella, indicating high variability of those genera in the human gut environment. Microdiversity analysis indicated that Ruminococcus bromii was statistically significantly (p < 0.002) more diverse on the strain level in the control samples compared to the PD samples. In addition, by clustering all genes and performing presence-absence analysis between groups, we identified several control-specific (p < 0.05) related genes, such as speF and Fe-S oxidoreductase. We also report detailed annotation of MAGs, including Clusters of Orthologous Genes (COG), Cas operon type, antiviral gene, prophage, and secondary metabolites biosynthetic gene clusters, which can be useful for providing a reference for future studies.

Keywords: Ruminococcus bromii; Metagenome; Microdiversity; Parkinson’s disease.

Figure 1

Maximum-likelihood phylogenetic tree including 943 dereplicated bacterial MAGs. Each tree branch represents a MAG. Bars on the outer layer represent how many MAGs were clustered within the dereplicated MAG, outer strip colours represent phylum of the MAGs. Background colour of the clade represents the annotated Cas type within the MAG.

Figure 2

Box-plot showing nucleotide diversity of Ruminococcus bromii in each sample. Each dot represents one sample. Blue box represents the control group, and the orange represents PD. The statistical significance was observed based on Wilcoxon rank-sum statistics and Benjamini/Hochberg correction. Samples with greater than 1× coverage depth were used for the statistical analysis.

Figure 3

Violinplot shows the Growth Rate Index (GRiD) score of each MAG grouped by (a) Phylum and (b) Family. Blue distribution diagram represents the control group, and the orange represents PD. White dot indicates the median. For simplicity, we only plotted the top eight family taxa with the highest number of GRiD record observations.

Figure 4

The pangenome of Prevotella MAGs. The dendrogram in the centre organises the 13,257 gene clusters that occur in all 64 Prevotella MAGs. The 64 inner circular layers correspond to the 64 Prevotella MAG. MAGs that were created from the control group and PD group are shown in blue and yellow, respectively. MAGs are ordered according to ANI scores which are shown at the top-right corner with red gradient. The length of the MAGs is shown with grey bars.

Figure 5

Box-plot illustrating the abundance of selected archaeal species across different samples. Each dot corresponds to a single sample. The blue box represents the control group, while the orange box represents the PD group. Panel (a) shows GUT_GENOME109037, and panel (b) shows GUT_GENOME247230, both of which are selected archaeal genomes from the human gut archaeome database.