Annotated bacterial chromosomes from frame-shift-corrected long-read metagenomic data

Abstract

Background: Short-read sequencing technologies have long been the work-horse of microbiome analysis. Continuing technological advances are making the application of long-read sequencing to metagenomic samples increasingly feasible.

Results: We demonstrate that whole bacterial chromosomes can be obtained from an enriched community, by application of MinION sequencing to a sample from an EBPR bioreactor, producing 6 Gb of sequence that assembles into multiple closed bacterial chromosomes. We provide a simple pipeline for processing such data, which includes a new approach to correcting erroneous frame-shifts.

Conclusions: Advances in long-read sequencing technology and corresponding algorithms will allow the routine extraction of whole chromosomes from environmental samples, providing a more detailed picture of individual members of a microbiome.

Keywords: Algorithms; Frame-shifts; Long-read sequencing; Microbial genomics; Microbiome; Sequence assembly; Software.

Fig. 1

Summary of results. a Bandage [30] visualization of the Unicycler assembly graph before final segmentation into contigs. The largest connected components are labeled by the corresponding taxonomic bins, and the nodes are colored by the MEGAN taxonomic classification of the corresponding long reads. The seven longest linear and circular components correspond to the seven LR-chromosomes. b MEGAN-LR taxonomic binning: nodes are scaled to indicate the number of aligned bases in each bin. Bins that are more than 50% complete are shown in bold. c Annotation of the seven LR-chromosomes, labeled by the corresponding taxonomic bins. The three circular tracks indicate the genes annotated by Prokka on the forward strand (blue) and reverse strand (pink), and GC-skew (green and red indicate lower or higher than average GC content, respectively)

Fig. 2

Analysis. a Long-read analysis pipeline shown from left to right. MinION sequencing produces a set of reads. These are assembled into contigs using Unicycler and aligned against the NCBI-nr database using DIAMOND. The contigs and alignments are processed by MEGAN so as to perform taxonomic binning and also to produce frame-shift-corrected contigs. These are analyzed using CheckM and annotated using Prokka. The duration of each step is shown in wall-clock hours. MEGAN analysis took less than 10 min. b Frame-shift correction: in frame-shift alignments, forward slashes, and backward slashes indicate a frame decrease, or increase, by one, respectively. Correction is performed by inserting one or two unspecified nucleotides into the sequence, respectively

Fig. 3

Dot plots for the three LR-chromosomes that have high similarity to reference genome assemblies, namely, B1 against GCA_001567185.1 (Bacteroidetes bacterium OLB12), B2 against GCA_000584975.2 (Candidatus accumulibacter sp. SK-02), and B5 against GCA_001567405.1 (Bacteroidetes bacterium OLB8). Forward alignments are shown in red, whereas reverse complemented alignments are shown in blue, and gray lines indicate contig boundaries in the reference assemblies. The number of contigs in each reference sequence is given in brackets

Fig. 4

Distribution of repeat rates in all complete bacterial genomes in RefSeq. Vertical lines show the repeat rate of the seven LR chromosomes. Additional 17 data points that have a repeat percentage above 25% are not shown

Fig. 5

For each of the seven LR chromosomes (B1–B7), we show a dot plot comparison against the set of SR contigs that align, reporting their number in brackets

Fig. 6

Overview of the concordance score for LR contigs, on the one hand, and SR-bins and reference genomes, on the other. The x axis shows the length of each LR contig, with the position of each tagged with a tick on the axis; the y axis shows the value of concordance score κ, and data points represent pairs of LR chromosomes and SR bins, or references genomes. Selected pairs with high concordance score are labeled with “ 〈LR-chromosome.id〉−〈SR-bin.id〉” for comparisons to SR-bins or “ 〈LR-chromosome.id〉−〈GCA_id〉” for comparisons to references. Within each set of the seven LR chromosome alignments, the pair with the maximum concordance score is shown in red. All LR chromosomes have highly concordant counterpart SR-bins, with the exception of LR chromosome 5. Further details on individual LR chromosomes are reported in Additional file 10: Figure S2