Search bioRxiv⌕ Search

bioRxiv · 10.64898/2026.11.09.765334

DNA methylation assigns sequencing reads to closely related bacterial strains when their sequence cannot

Abstract

Clinical and public-health questions often concern strains rather than species, yet co-resident strains of one species can be so closely related that a single long read carries no sequence difference between them. We measured how often this happens in 639 pairs of isolates of Enterococcus faecium, Klebsiella pneumoniae, Listeria monocytogenes and Staphylococcus aureus sequenced on one nanopore platform. Among the 297 pairs above 99% average nucleotide identity (ANI), 40.37% of read placements (a read counted against one strain's genome) carried no sequence difference between the two strains, and above 99.9% ANI about three quarters did. Strains of one species also differ in the DNA motifs they methylate. Our method, Epilect, assigns single nanopore reads to strains from the basecaller's methylation calls, given a methylation profile for each strain, and declines to call when the evidence is weak. In the pairs above 99% ANI it correctly assigned 21.8% of the read placements that sequence could not place, at a precision of 0.959. In computational mixtures of 65 pairs of distinct strains at or above 99.5% ANI that were predicted to differ in methylation, its per-read F1 score exceeded that of the best of three sequence-based read-phasing tools by 0.25 (95% CI 0.08 to 0.37); where equally related strains were predicted to share a methylome, the comparison reversed. DNA methylation therefore supplies strain evidence on reads whose sequence carries none, and whether it will do so for a given pair of strains can be estimated before sequencing.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Kakabra, K., Claesson, M. J.. 2026-11-15. DNA methylation assigns sequencing reads to closely related bacterial strains when their sequence cannot. https://doi.org/10.64898/2026.11.09.765334

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

Point mutations in the FimH lectin domain: predicted folding stability versus conformational dynamics for R60P, Q133K and Y137A

Background. The adhesin FimH attaches uropathogenic Escherichia coli to the urinary epithelium through a force-regulated catch-bond that single lectin-domain substitutions can modulate. How such substitutions reshape the dynamics of the isolated lectin domain, and whether stability predictors anticipate them, is unclear. Methods. We compared the wild-type (WT) lectin domain (PDB 4XO8, apo) with R60P, Y137A and Q133K, using three 500 ns molecular dynamics (MD) replicates per system (6 microseconds total) and three folding-stability predictors (FoldX, a Cartesian-ddG-type Rosetta protocol, ERIS-2). The replicate was the unit of analysis (n = 3), with 95% confidence intervals and a WT replicate-to-replicate control. Results. FoldX and ERIS-2 ranked the variants consistently, R60P most destabilising and Q133K neutral; the Rosetta value for Q133K was anomalously high. All systems stayed folded. R60P gave the only reproduced local change, loss of beta-strand structure at residues 60-61 and of three local hydrogen bonds. Y137A did not bring Thr158 into the pocket, and the Q133K Gln133-Asp54 polar contact became a Lys133-Asp54 salt bridge. In the principal-component landscape, departures from the WT region came from individual replicates (Y137A replicate 1; Q133K replicates 2-3), not a systematic shift. Conclusions. A substitution perturbs local structure most clearly when it breaks a beta-strand (R60P), and the stability predictions and dynamics together give a consistent ranking of the variants, including the first dynamic characterisation of Q133K. The results characterise the apo lectin domain and give a reproducible basis for interpreting FimH mutations, not for inferring affinity. Keywords: FimH; lectin domain; point mutations; molecular dynamics; folding stability ({Delta}{Delta}G); protein stability prediction; principal component analysis; urinary tract infection.

bioinformatics↗

Inferring single-cell gene expression profiles from multicellular spatial transcriptomics data with st2sc

Spatial transcriptomics (ST) enables the quantitative measurement of gene expression within tissue sections but is currently limited in spatial resolution, as each spot aggregates signals from multiple cells. Recovering single-cell gene expression profiles from these multicellular measurements requires two steps: explicit identification of individual cells and accurate estimation of per-gene expression. While existing methods have demonstrated the feasibility of recovering single-cell expression profiles, accurately estimating gene-level expression for individual cells remains an important challenge. To address this challenge, we propose st2sc (Spatial Transcriptomics to Single-Cell), an end-to-end framework for inferring single-cell gene profiles from multicellular ST data. st2sc introduces a novel Attention Gradient Flow Tracking (AGFT) algorithm that extracts attention maps from a Vision-Gene Transformer to identify candidate cell instances without predefined background constraints. The detected instances are then used by a set of regressors trained under a multiple-instance, multiple-target regression setting to predict precise single-cell gene expression counts. Comprehensive evaluations on synthetic and Visium-like datasets demonstrate that st2sc significantly outperforms state-of-the-art methods in accurately segmenting cellular instances and predicting gene counts. The method can further resolve contested 2um bin in Visium HD. Application to human IPF lung showed the model can recover the subtype marker level analysis. This advancement further improves single-cell resolution analysis in spatial transcriptomics, facilitating more accurate downstream biological analyses.

bioinformatics↗

From Screening to Differential Diagnosis: Benchmarking 3D Facial Representations Across Craniofacial Conditions

Rare genetic disorders often present with highly heterogeneous physical features: some have distinctive craniofacial characteristics, whereas others produce subtler or more variable facial differences, making phenotype-based diagnosis challenging. Three-dimensional (3D) facial scanning is a noninvasive approach that captures facial shape and surface geometry, providing quantitative information on craniofacial morphology beyond conventional 2D photographs. Here, we investigated how well 3D facial representations support clinical tasks ranging from broad screening to fine-grained differential diagnosis without landmark registration, dense correspondence, or template fitting. We benchmarked four 3D models, including PointNet, PointNet++, DGCNN, and a geometric baseline, using 26,990 facial scans from 18,785 subjects across four prediction tasks. Across models, performance declined as greater diagnostic specificity was required. Broad affected-versus-unaffected screening achieved high performance (accuracy 0.893; AUC 0.956), although sensitivity analyses revealed substantial influences of cohort composition and demographic differences. Fine-grained syndrome diagnosis remained more challenging, with limited top-1 accuracy despite meaningful class-wise discrimination. Combining repeated scans improved patient-level diagnosis beyond differences in patient weighting. Prototype inference further improved diagnostic prioritization, placing the correct syndrome among the top five candidates for about 65\% of held-out patients across 49 diagnoses. Patient-to-patient retrieval identified phenotypically similar confirmed cases, while useful syndrome prototypes could be constructed from relatively few reference patients. Anatomical region-occlusion analysis revealed syndrome-dependent patterns of facial importance broadly consistent with clinically recognized craniofacial features, although overlapping facial phenotypes contributed to diagnostic confusion. Overall, these findings support the potential of 3D facial representations for diagnostic candidate prioritization and retrieval, while emphasizing the influence of cohort composition on screening performance and the challenges of distinguishing phenotypically similar genetic syndromes.

bioinformatics↗