David A. Eccles

Vaccines capable of eliciting broadly neutralizing antibodies (bnAbs) are a major goal for pandemic preparedness. A challenge across vaccine fields is how to deliberately recruit the rare B cell clones that recognize conserved epitopes shared across diverse viral variants. BnAbs have been known to frequently emerge through extensive somatic hypermutation during affinity maturation, here we describe an alternative mechanism for bnAb selection. We designed an mRNA vaccine in which two antigenically distinct (severe acute respiratory syndrome coronavirus 2) SARS-CoV-2 variant’s (Omicron and Delta; O-Δ) receptor binding domains (RBDs) are physically fused on a single polypeptide. This design is predicted to favor B cell antigen receptors capable of engaging conserved epitopes on both RBDs with enhanced avidity. A matched nondivergent tandem RBD (Delta-Delta; Δ-Δ) served as a control. In mice, immunization elicited potent antibody responses and increased the frequency of cross-reactive B cells, recognizing Delta, Omicron, and the 2002 pandemic strain SARS-CoV RBDs. Using multicolor RBD tetramers and single-cell B cell receptor sequencing, we show that breadth arises via two distinct pathways. The divergent vaccine preferentially enriches clonally distinct cross-reactive B cells (not present within non-cross-reactive B cell pools) with low levels of somatic hypermutation, consistent with selection of germline-biased precursors. In contrast, the matched control vaccine yields cross-reactivity primarily within existing clonal lineages (clonal overlap between cross-reactive and non-cross-reactive cells) and at higher mutational burdens, consistent with affinity-maturation-driven acquisition of breadth. This work establishes a simple, modular antigen-design principle in which juxtaposing appropriately divergent antigens on a single scaffold promotes the enrichment of bnAb-prone B cells.

Immune competent animal models are essential in preclinical glioma research. The ability to investigate tumor development with key tumor microenvironment components such as immune infiltration, stromal cells and extracellular matrix allows for the investigation of these complex tumors. However, the current range of syngeneic models of glioma possess intrinsic limitations which must be acknowledged when designing a preclinical study. To address this gap, we developed genetically engineered mouse cell line models (GEM-CLeMs) by introducing common glioma driver mutations into immortalized astrocytes. A high-grade glioma model was generated by combining Pten knockdown with RAS V12 overexpression, while a low-grade glioma model was produced through p53 knockdown with mutant IDH1 ^R132H overexpression. The RAS/Pten GEM-CLeM tumors grew rapidly in vivo , displayed necrosis, multinucleated pleomorphisms, abundant vascularization, and showed strong enrichment for extracellular matrix remodelling and mesenchymal genes, features closely aligned with human glioblastoma. Importantly, the RAS/Pten GEM-CLeM tumors showed similar survival trends and immune infiltration patterns as a corresponding Kras ^G12D / Pten ^cKO GEMM, but could be grown to larger sizes, facilitating better stromal and immune analyses. In contrast, the IDH1 ^R132H /p53 GEM-CLeM formed slow-growing tumors with distinctive immune infiltration and vascular patterns, consistent with low-grade glioma phenotypes. Compared with the commonly used cell line GL261, the GEM-CLeM tumors had higher levels of stromal integration and immune suppression, making them a more faithful model of the glioma tumor microenvironment. This system enables rapid generation of transplantable glioma models with defined driver mutations in a low-mutational background, offering a flexible platform for dissecting glioma biology and evaluating immunotherapies. Importance of the study: We have generated a set of customizable, modular Genetically Engineered Mouse Cell Line Models (GEM-CLeMs) of “high grade” and “low grade” glioma. They reliably form tumors when transplanted intracranially into immune-competent C57BL/6 mice, and they are cost- and time-effective at capturing the important characteristics of glioma, both mutant IDH1 low grade glioma and high grade glioblastoma. Critically these characteristics include the myeloid-rich immune suppressive tumor immune microenvironment, a key weakness of existing murine glioma cell lines like GL261. These GEM-CLeM models can be used in multiple ways. The cells are amenable to further manipulation, so therapeutic targets and drug mechanism of action can be assessed. The activity of candidate genes in tumour formation and phenotype can be determined. Most importantly the models can be used to develop effective immunotherapies, including strategies to target macrophage and myeloid cell immune suppression. Key points: Combinations of driver mutations were engineered into an immortalised mouse astrocyte. Engineered cells formed tumours on intracranial transplant into immune competent mice. Tumors had key histological and immune suppressive features of human glioma.

This protocol demonstrates how to assemble reads from plasmid DNA, and generate a circularised and non-repetitive consensus sequence. For plasmid sample preparation for nanopore sequencing, see here. The most recent version of our rapid barcoding kit protocol at the time this protocol was written is attached below: At the moment, this protocol uses Canu to de-novo assemble high-quality single-cut reads. Input(s): demultiplexed fastq files (see protocol Demultiplexing Nanopore reads with LAST). I've noticed that the default demultiplexing carried out by Guppy and Dorado has some issues with chimeric reads, which can affect assembly. Output(s): Consensus sequence per barcode as a fasta file

The ongoing degradation of natural systems and other environmental changes has put our society at a crossroad with respect to our future relationship with our planet. While the concept of One Health describes how human health is inextricably linked with environmental health, many of these complex interdependencies are still not well‐understood. Here, we describe how the advent of real‐time genomic analyses can benefit One Health and how it can enable timely, in‐depth ecosystem health assessments. We introduce nanopore sequencing as the only disruptive technology that currently allows for real‐time genomic analyses and that is already being used worldwide to improve the accessibility and versatility of genomic sequencing. We showcase real‐time genomic studies on zoonotic disease, food security, environmental microbiome, emerging pathogens, and their antimicrobial resistances, and on environmental health itself – from genomic resource creation for wildlife conservation to the monitoring of biodiversity, invasive species, and wildlife trafficking. We stress why equitable access to real‐time genomics in the context of One Health will be paramount and discuss related practical, legal, and ethical limitations.

This protocol demonstrates how to assemble reads from plasmid DNA, and generate a circularised and non-repetitive consensus sequence At the moment, this protocol uses Canu to de-novo assemble high-quality single-cut reads. Input(s): demultiplexed fastq files (see protocol Demultiplexing Nanopore reads with LAST). I've noticed that the default demultiplexing carried out by Guppy (at least up to v4.2.2, as used in the first version of this protocol) has issues with chimeric reads, which can affect assembly. Output(s): Consensus sequence per barcode as a fasta file

This protocol is for preparing long reads for stranded mapping, as an intermediate step for additional protocols: Aligning strand-oriented sequences to a transcriptome for transcript / gene counting Aligning strand-oriented sequences to a genome for confirmatory QC Input(s): demultiplexed fastq files (see protocol Demultiplexing Nanopore reads with LAST), adapter file (containing strand-sensitive adapter sequences) Output(s): oriented read files, as gzipped fastq files

This protocol is for a semi-manual method for read demultiplexing, as used after my presentation Sequencing DNA with Linux Cores and Nanopores to work out the number of reads captured by different barcodes. This approach has now been bundled up into a single Perl script, fastq-dental.pl which can be found here. Input: reads as a FASTQ file, barcode sequences as a FASTA file Output: reads split into single FASTQ files per target [barcode] Note: barcode / adapter sequences are not trimmed by this protocol. This allows for subsequent adapter-associated downstream processing (e.g. strand correction for cDNA reads).

This protocol is for generating count tables for different samples at the transcript level, using long reads that are mapped to transcripts. Input(s): demultiplexed and oriented fastq files (see protocol Preparing Reads for Stranded Mapping), transcript reference fasta file, annotation file Output(s): transcript table, sorted by summed coverage across all samples, annotated with gene name / description / location

This protocol is for generating count tables for different samples at the transcript level, using long reads that are mapped to transcripts. Input(s): demultiplexed and oriented fastq files (see protocol Preparing Reads for Stranded Mapping), transcript reference fasta file, annotation file Output(s): transcript table, sorted by summed coverage across all samples, annotated with gene name / description / location