Palidis: fast discovery of novel insertion sequences

Abstract

The diversity of microbial insertion sequences, crucial mobile genetic elements in generating diversity in microbial genomes, needs to be better represented in current microbial databases. Identification of these sequences in microbiome communities presents some significant problems that have led to their underrepresentation. Here, we present a bioinformatics pipeline called Palidis that recognizes insertion sequences in metagenomic sequence data rapidly by identifying inverted terminal repeat regions from mixed microbial community genomes. Applying Palidis to 264 human metagenomes identifies 879 unique insertion sequences, with 519 being novel and not previously characterized. Querying this catalogue against a large database of isolate genomes reveals evidence of horizontal gene transfer events across bacterial classes. We will continue to apply this tool more widely, building the Insertion Sequence Catalogue, a valuable resource for researchers wishing to query their microbial genomes for insertion sequences.

Keywords: antimicrobial resistance; horizontal gene transfer; insertion sequences; metagenome; mobile genetic element; transposon.

Fig. 1.

Steps summarizing Palidis. Step 1: reads from mixed microbial communities are pre-processed and run through pal-MEM to identify reads containing repeat sequences. Step 2: reads containing repeat sequences are mapped against the assemblies using Bowtie2 to find their positions and proximity filters applied to identify candidate ITRs. Step 3: candidate ITRs are clustered using CD-HIT-EST. ISs are identified by ITRs that are of the same cluster and are reverse complements of each other. Step 4: search of transposases using InterProScan. Step 5: final outputs of a FASTA file with insertion sequences and tab-delimited file with information are created.

Fig. 2.

(a) Length of insertion sequences and (b) number of transposases within the ISC .

Fig. 3.

Number of samples from the 661 k database that contain an IS found in across (a) genera; (b) species in the 661 k database a.

Fig. 4.

(a) Number of ISs shared across a number of distinct, known genera in the 661 k database; (b) network of shared ISs between genera. Each vertex represents either an IS (in ISfinder: blue or not in ISfinder: read) or a genus (labelled); (c) Number of ISs that are either in ISfinder (blue) or not (red) that are in distinct, known species in the 661 k database.