Soeren Michael Karst

Machine learning holds great promise for accelerating enzyme optimization, but its power is fundamentally constrained by the limited availability of sequence–fitness data. Here, we introduce MillionFull , a low-cost method that enables high-throughput full-length sequence– fitness mapping for enzymes of arbitrary length. Each run yields on the order of 10⁵–10⁷ data points, capturing sequence–function relationships at unprecedented scale. By overcoming the data bottleneck, MillionFull provides a foundation for dramatically advancing AI-driven enzyme engineering.

Background Bloodstream infections (BSIs) remain a major cause of mortality, in part due to many patients developing sepsis or septic shock. To survive sepsis, it is paramount that effective antimicrobial therapy is initiated rapidly to avoid excess mortality, but the current gold-standard to identify the pathogen in BSIs, blood culturing, has great limitations with a long turnaround time and a poor sensitivity. This delay to correct empiric broad-spectrum antimicrobial treatments leads to excess mortality and antimicrobial resistance development. Methods In this study we developed a metagenomic next-generation sequencing (mNGS) assay utilizing the Oxford Nanopore Technologies platform to sequence microbial cell-free DNA from blood plasma. The method was evaluated in a prospective observational clinical study (n=40) in an emergency ward setting, where a study sample was taken from the same venipuncture as a blood culture sample from patients with a suspected BSI. Findings Nanopore mNGS confirmed all findings in patients with a positive blood culture (n=11), and identified pathogens relevant to the acute infection in an additional 11 patients with a negative blood culture. In an analysis of potential impact on the antibiotic treatment, we found that 59% (n=13) of mNGS positive answers could have impacted the treatment, with five cases of a change from ineffective to effective therapy. Interpretation This study demonstrates that culture-independent Nanopore mNGS directly on blood plasma could be a feasible alternative to blood culturing for infection diagnostics for patients admitted with a severe infection or sepsis. The method identified a relevant pathogen in patients with a broad range of etiologies including urinary tract infections and lower respiratory tract infections. With a turnaround time of 6 hours the method could provide unprecedented speed and sensitivity in BSI diagnostics.

Reductions in sequencing costs have enabled widespread use of shotgun metagenomics and amplicon sequencing, which have drastically improved our understanding of the microbial world. However, large sequencing projects are now hampered by the cost of library preparation and low sample throughput. Here, we benchmarked three high-throughput DNA extraction methods: ZymoBIOMICS™ 96 MagBead DNA Kit, MP Biomedicals ^TM FastDNA ^TM -96 Soil Microbe DNA Kit, and DNeasy® 96 PowerSoil® Pro QIAcube® HT Kit. The DNA extractions were evaluated based on length, quality, quantity, and the observed microbial community across five diverse soil types. DNA extraction of all soil types was successful for all kits, however DNeasy® 96 PowerSoil® Pro QIAcube® HT Kit excelled across all performance parameters. We further used the nanoliter dispensing robot I.DOT One to miniaturize Illumina amplicon and metagenomic library preparation volumes by a factor of 5 and 10, respectively, with no significant impact on the observed microbial communities. With these protocols, DNA extraction, metagenomic library preparation, or amplicon library preparation for one 96-well plate are approx. 3, 5, and 6 hours, respectively. Furthermore, the miniaturization of amplicon and metagenome library preparation reduces the chemical and plastic costs from 5.0 to 3.6 and 59 to 7.3 USD pr. sample.

16S rRNA and shotgun metagenomics studies typically yield different results, usually attributed to biases in PCR amplification of 16S rRNA genes. Here, we introduce Greengenes2 and show that differences in reference phylogeny are more important. By inserting sequences into a whole-genome phylogeny, we show that 16S rRNA and shotgun metagenomic data generated from the same samples agree in principal coordinates space, taxonomy, and in phenotype effect size when analyzed with the same tree.

Long-read Oxford Nanopore sequencing has democratized microbial genome sequencing and enables the recovery of highly contiguous microbial genomes from isolates or metagenomes. However, to obtain near-finished genomes it has been necessary to include short-read polishing to correct insertions and deletions derived from homopolymer regions. Here, we show that Oxford Nanopore R10.4 can be used to generate near-finished microbial genomes from isolates or metagenomes without short-read or reference polishing.

Long-read Oxford Nanopore sequencing has democratized microbial genome sequencing and enables the recovery of highly contiguous microbial genomes from isolates or metagenomes. However, to obtain near-perfect genomes it has been necessary to include short-read polishing to correct insertions and deletions derived from homopolymer regions. Here, we show that Oxford Nanopore R10.4 can be used to generate near-perfect microbial genomes from isolates or metagenomes without shortread or reference polishing.

High-throughput amplicon sequencing of large genomic regions remains challenging for short-read technologies. Here, we report a high-throughput amplicon sequencing approach combining unique molecular identifiers (UMIs) with Oxford Nanopore Technologies (ONT) or Pacific Biosciences circular consensus sequencing, yielding high-accuracy single-molecule consensus sequences of large genomic regions. We applied our approach to sequence ribosomal RNA operon amplicons (~4,500 bp) and genomic sequences (>10,000 bp) of reference microbial communities in which we observed a chimera rate <0.02%. To reach a mean UMI consensus error rate <0.01%, a UMI read coverage of 15× (ONT R10.3), 25× (ONT R9.4.1) and 3× (Pacific Biosciences circular consensus sequencing) is needed, which provides a mean error rate of 0.0042%, 0.0041% and 0.0007%, respectively.

High-throughput 16S rRNA gene amplicon sequencing is an essential method for studying the diversity and dynamics of microbial communities. However, this method is presently hampered by the lack of high-identity reference sequences for many environmental microbes in the public 16S rRNA gene reference databases and by the absence of a systematic and comprehensive taxonomy for the uncultured majority. Here, we demonstrate how high-throughput synthetic long-read sequencing can be applied to create ecosystem-specific full-length 16S rRNA gene amplicon sequence variant (FL-ASV) resolved reference databases that include high-identity references (>98.7% identity) for nearly all abundant bacteria (>0.01% relative abundance) using Danish wastewater treatment systems and anaerobic digesters as an example. In addition, we introduce a novel sequence identity-based approach for automated taxonomy assignment (AutoTax) that provides a complete seven-rank taxonomy for all reference sequences, using the SILVA taxonomy as a backbone, with stable placeholder names for unclassified taxa. The FL-ASVs are perfectly suited for the evaluation of taxonomic resolution and bias associated with primers commonly used for amplicon sequencing, allowing researchers to choose those that are ideal for their ecosystem. Reference databases processed with AutoTax greatly improves the classification of short-read 16S rRNA ASVs at the genus- and species-level, compared with the commonly used universal reference databases. Importantly, the placeholder names provide a way to explore the unclassified environmental taxa at different taxonomic ranks, which in combination with in situ analyses can be used to uncover their ecological roles.

Microorganisms are critical to water recycling, pollution removal and resource recovery processes in the wastewater industry. While the structure of this complex community is increasingly understood based on 16S rRNA gene studies, this structure cannot currently be linked to functional potential due to the absence of high-quality metagenome-assembled genomes (MAGs) with full-length rRNA genes for nearly all species. Here, we sequence 23 Danish full-scale wastewater treatment plant metagenomes, producing >1 Tbp of long-read and >0.9 Tbp of short-read data. We recovered 1083 high-quality MAGs, including 57 closed circular genomes. The MAGs accounted for ~30% of the community, and meet the stringent MIMAG high-quality draft requirements including full-length rRNA genes. We show how novel high-quality MAGs in combination with >13 years of amplicon data, Raman microspectroscopy and fluorescence in situ hybridisation can be used to uncover abundant undescribed lineages belonging to important functional groups.

High-throughput 16S rRNA gene amplicon sequencing is an essential method for studying the diversity and dynamics of microbial communities. However, this method is presently hampered by the lack of high-identity reference sequences for many environmental microbes in the public 16S rRNA gene reference databases, and by the absence of a systematic and comprehensive taxonomy for the uncultured majority. Here we demonstrate how high-throughput synthetic long-read sequencing can be applied to create ecosystem-specific full-length 16S rRNA gene amplicon sequence variant (FL-ASV) reference databases that include high-identity references (>98.7% identity) for nearly all abundant bacteria (>0.01% relative abundance) using Danish wastewater treatment systems and anaerobic digesters as an example. In addition, we introduce a novel sequence identity-based approach for automated taxonomy assignment (AutoTax) that provides a complete seven-rank taxonomy for all reference sequences, using the SILVA taxonomy as a backbone, with stable placeholder names for unclassified taxa. The FL-ASVs are perfectly suited for the evaluation of taxonomic resolution and bias associated with primers commonly used for amplicon sequencing, allowing researchers to choose those that are ideal for their ecosystem. The AutoTax taxonomy greatly improves the classification of short-read 16S rRNA gene amplicon sequence variants (ASVs) at the genus- and species-level, compared to the commonly used universal reference databases. Importantly, the placeholder names provide a way to explore the unclassified environmental taxa at different taxonomic ranks, which in combination with in situ analyses can be used to uncover their ecological roles.