Utilizing co-abundances of antimicrobial resistance genes to identify potential co-selection in the resistome.

The rapid spread of antimicrobial resistance (AMR) is a threat to global health, and the nature of co-occurring antimicrobial resistance genes (ARGs) may cause collateral AMR effects once antimicrobial agents are used. Therefore, it is essential to identify which pairs of ARGs co-occur. Given the wealth of next-generation sequencing data available in public repositories, we have investigated the correlation between ARG abundances in a collection of 214,095 metagenomic data sets. Using more than 6.76∙108 read fragments aligned to acquired ARGs to infer pairwise correlation coefficients, we found that more ARGs correlated with each other in human and animal sampling origins than in soil and water environments. Furthermore, we argued that the correlations could serve as risk profiles of resistance co-occurring to critically important antimicrobials (CIAs). Using these profiles, we found evidence of several ARGs conferring resistance for CIAs being co-abundant, such as tetracycline ARGs correlating with most other forms of resistance. In conclusion, this study highlights the important ARG players indirectly involved in shaping the resistomes of various environments that can serve as monitoring targets in AMR surveillance programs. IMPORTANCE: Understanding the collateral effects happening in a resistome can reveal previously unknown links between antimicrobial resistance genes (ARGs). Through the analysis of pairwise ARG abundances in 214K metagenomic samples, we observed that the co-abundance is highly dependent on the environmental context and argue that these correlations can be used to show the risk of co-selection occurring in different settings.

ARGprofiler-a pipeline for large-scale analysis of antimicrobial resistance genes and their flanking regions in metagenomic datasets.

MOTIVATION: Analyzing metagenomic data can be highly valuable for understanding the function and distribution of antimicrobial resistance genes (ARGs). However, there is a need for standardized and reproducible workflows to ensure the comparability of studies, as the current options involve various tools and reference databases, each designed with a specific purpose in mind. RESULTS: In this work, we have created the workflow ARGprofiler to process large amounts of raw sequencing reads for studying the composition, distribution, and function of ARGs. ARGprofiler tackles the challenge of deciding which reference database to use by providing the PanRes database of 14 078 unique ARGs that combines several existing collections into one. Our pipeline is designed to not only produce abundance tables of genes and microbes but also to reconstruct the flanking regions of ARGs with ARGextender. ARGextender is a bioinformatic approach combining KMA and SPAdes to recruit reads for a targeted de novo assembly. While our aim is on ARGs, the pipeline also creates Mash sketches for fast searching and comparisons of sequencing runs. AVAILABILITY AND IMPLEMENTATION: The ARGprofiler pipeline is a Snakemake workflow that supports the reuse of metagenomic sequencing data and is easily installable and maintained at https://github.com/genomicepidemiology/ARGprofiler.

A curated data resource of 214K metagenomes for characterization of the global antimicrobial resistome.

The growing threat of antimicrobial resistance (AMR) calls for new epidemiological surveillance methods, as well as a deeper understanding of how antimicrobial resistance genes (ARGs) have been transmitted around the world. The large pool of sequencing data available in public repositories provides an excellent resource for monitoring the temporal and spatial dissemination of AMR in different ecological settings. However, only a limited number of research groups globally have the computational resources to analyze such data. We retrieved 442 Tbp of sequencing reads from 214,095 metagenomic samples from the European Nucleotide Archive (ENA) and aligned them using a uniform approach against ARGs and 16S/18S rRNA genes. Here, we present the results of this extensive computational analysis and share the counts of reads aligned. Over 6.76∙108 read fragments were assigned to ARGs and 3.21∙109 to rRNA genes, where we observed distinct differences in both the abundance of ARGs and the link between microbiome and resistome compositions across various sampling types. This collection is another step towards establishing global surveillance of AMR and can serve as a resource for further research into the environmental spread and dynamic changes of ARGs.

Global Distribution of mcr Gene Variants in 214K Metagenomic Samples.

Since the initial discovery of a mobilized colistin resistance gene (mcr-1), several other variants have been reported, some of which might have circulated a while beforehand. Publicly available metagenomic data provide an opportunity to reanalyze samples to understand the evolutionary history of recently discovered antimicrobial resistance genes (ARGs). Here, we present a large-scale metagenomic study of 442 Tbp of sequencing reads from 214,095 samples to describe the dissemination and emergence of nine mcr gene variants (mcr-1 to mcr-9). Our results show that the dissemination of each variant is not uniform. Instead, the source and location play a role in the spread. However, the genomic context and the genes themselves remain primarily unchanged. We report evidence of new subvariants occurring in specific environments, such as a highly prevalent and new variant of mcr-9. This work emphasizes the importance of sharing genomic data for the surveillance of ARGs in our understanding of antimicrobial resistance. IMPORTANCE The ever-growing collection of metagenomic samples available in public data repositories has the potential to reveal new details on the emergence and dissemination of mobilized colistin resistance genes. Our analysis of metagenomes deposited online in the last 10 years shows that the environmental distribution of mcr gene variants depends on sampling source and location, possibly leading to the emergence of new variants, although the contig on which the mcr genes were found remained consistent.

NetSolP: predicting protein solubility in Escherichia coli using language models.

MOTIVATION: Solubility and expression levels of proteins can be a limiting factor for large-scale studies and industrial production. By determining the solubility and expression directly from the protein sequence, the success rate of wet-lab experiments can be increased. RESULTS: In this study, we focus on predicting the solubility and usability for purification of proteins expressed in Escherichia coli directly from the sequence. Our model NetSolP is based on deep learning protein language models called transformers and we show that it achieves state-of-the-art performance and improves extrapolation across datasets. As we find current methods are built on biased datasets, we curate existing datasets by using strict sequence-identity partitioning and ensure that there is minimal bias in the sequences. AVAILABILITY AND IMPLEMENTATION: The predictor and data are available at https://services.healthtech.dtu.dk/service.php?NetSolP and the open-sourced code is available at https://github.com/tvinet/NetSolP-1.0. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Deep protein representations enable recombinant protein expression prediction.

A crucial process in the production of industrial enzymes is recombinant gene expression, which aims to induce enzyme overexpression of the genes in a host microbe. Current approaches for securing overexpression rely on molecular tools such as adjusting the recombinant expression vector, adjusting cultivation conditions, or performing codon optimizations. However, such strategies are time-consuming, and an alternative strategy would be to select genes for better compatibility with the recombinant host. Several methods for predicting soluble expression are available; however, they are all optimized for the expression host Escherichia coli and do not consider the possibility of an expressed protein not being soluble. We show that these tools are not suited for predicting expression potential in the industrially important host Bacillus subtilis. Instead, we build a B. subtilis-specific machine learning model for expressibility prediction. Given millions of unlabelled proteins and a small labeled dataset, we can successfully train such a predictive model. The unlabeled proteins provide a performance boost relative to using amino acid frequencies of the labeled proteins as input. On average, we obtain a modest performance of 0.64 area-under-the-curve (AUC) and 0.2 Matthews correlation coefficient (MCC). However, we find that this is sufficient for the prioritization of expression candidates for high-throughput studies. Moreover, the predicted class probabilities are correlated with expression levels. A number of features related to protein expression, including base frequencies and solubility, are captured by the model.