BGPT: Paper Review: Harnessing Diversity Generating Retroelements forin vivotargeted hyper-mutagenesis

Explore by Goal

Quick Explanation Copied

DGRec (DGR + ssDNA recombineering) in E. coli enables programmable, in vivo, targeted adenine-biased hypermutagenesis, with quantified RT context-dependence and demonstrated phage/bacterial receptor “arms-race” functionality. Evidence is directly supported by the authors’ sequencing-based mutation profiling, controls, and SRA-deposited datasets for this work. .

Long Explanation

Paper Review (critical, visual, evidence-based): DGRec for in vivo targeted hyper-mutagenesis

Target paper: “Harnessing Diversity Generating Retroelements for in vivo targeted hyper-mutagenesis” (10.1101/2025.03.24.644984) .

1) Visual map of the DGRec mechanism (what they built)

Evidence basis: The paper’s abstract + results explicitly describe DGR RT bias on TR, and DGRec coupling to recombineering via cspRecT and mutL* to enable overwriting/recombination at a targeted VR region in E. coli. .

2) Quantitative mutation profiling (what the data show)

The authors report UMI-corrected amplicon sequencing for a sacB target TR. They highlight a strongly elevated mutation rate on adenines within the targeted TR/VR region compared to outside-region background, and they report additional context effects and multi-mutation co-occurrence. .

Skeptical note: the “2+ mutations” figure is reported as a fraction among mutated molecules, not as an absolute probability per target molecule without selection; interpreting this into a generative model requires additional raw distribution details. .

3) Context dependence: immediate bases bias adenine-targeted mutagenesis

The authors report that mutation profiles are reproducible per engineered TR and that local sequence context influences mutation rates and nucleotide incorporation biases for adenines (including effects of +1/+2 positions and “snowball” behavior). .

Uncertainty / limitation: the excerpted text provides qualitative and some quantitative context effects but not a full numerical table for each context combination in the provided portion, so the plot cannot reconstruct the complete context map without additional raw figures/tables. .

4) Programmable target diversification & functional phage/bacteria “arms race”

The core functional claim is that DGRec can diversify phage receptor-binding protein GpJ (λvir) and bacterial LamB receptor variants, yielding reciprocal escape/reinfection dynamics. The paper describes: (i) diversifying LamB loops then selecting λvir-resistant clones (including complete and partial resistance fractions), (ii) using a DGRec-constructed phage GpJ library to reinfect those LamB variants, and (iii) observing selection-enriched GpJ substitutions at the receptor-binding interface. .

Skeptical check: the reported fractions are from phenotypic plaques/zone readouts; the excerpt does not provide whether partial resistance clones still display LamB expression/function quantitatively. Strengthening causality would require orthogonal confirmation (e.g., direct LamB structural/functional readouts) beyond the sequencing described. .

5) Controls, modularity, and what could break

Key positive/negative controls reported

Inactive bRT abolishes mutagenesis and sucrose-resistance signal in the sacB assay; removing Avd also abolishes mutagenesis. .
No CspRecT eliminates mutagenesis, supporting that recombination step is required and endogenous E. coli pathways do not substitute. .
Some recombination-related deletions (sbcB, recJ) increase efficiency, suggesting recombineering background conditions matter. .

Known limitation acknowledged by authors

The authors explicitly note that high levels of mutagenesis could not be obtained with all TR sequences, and they propose future work to identify sequence features that impact DGRec efficiency. .

Biological failure modes (inference; what the data implies)

Recombination bottlenecks: since mutagenesis depends on recombineering capability (cspRecT + mutL* and recombination-promoting genomic background), varying genomic context could change efficiency. (Inference from control/modularity description; not directly quantified for every context.) .
Sequence-context design constraints: adenine-biased mutagenesis is programmable via TR, but RT biases (including template U effects and codon-level outcomes) mean not all amino-acid landscapes are explored uniformly. .
Multi-mutation spectra: frequent 2+ mutation molecules might be beneficial for exploring epistasis, but it complicates genotype-to-phenotype attribution if downstream selection is performed without cloning/individualization. .

6) Relevance to the wider DGR/retron field (grounded context)

DGRec is conceptually aligned with the broader idea that DGRs are natural systems producing targeted diversity using RT bias patterns (adenine-directed) and TR/VR re-writing. .

More broadly, the paper sits inside synthetic-biology momentum to create in vivo genome editing / continuous evolution tools and uses modular components and sequencing readouts to map mutational landscapes. .

7) Data & code transparency (replication-critical)

Sequencing data: deposited in SRA under Bioproject PRJNA1140560. .
Code: custom Python package for analysis is provided at https://github.com/dbikard/dgrec, with analysis notebooks at https://gitlab.pasteur.fr/dbikard/dgrec_analysis. .

8) What would most efficiently disprove/reshape the paper’s core claims?

Mechanistic causal test of recombineering dependence: demonstrate that mutation/recombination at the target VR fails when recombineering components are removed or when repair suppression is disabled, beyond the controls already shown—ideally across multiple loci/TRs to show generality rather than one sacB case. .
Quantitative genotype-to-phenotype mapping: the functional arms-race conclusion could be strengthened by cloning individual selected genotypes to separate linked multi-mutation effects and confirm that the specific enriched substitutions are sufficient for escape/reinfection. .

Author Review Links (bespoke)

Feedback:

Updated: April 28, 2026