DNA-PAINT enables super-resolution imaging with high multiplexing capacity, yet it is rarely used in nanoscale bacteriology. A likely reason is the lack of straightforward protocols and of systematic characterization of imager-docking strand pairs inside intact bacterial cells, whose envelope and dense cytoplasm differ markedly from both the DNA origami in vitro systems used to develop them, and eukaryotic cells. In this work, we establish a screening platform based on RNA-FISH, in which we extend the 16S rRNA-directed EUB338 probe with docking sequences of choice. This allows imager-docking pairs to be benchmarked directly inside Escherichia coli cells and characterized for parameters such as specificity, binding kinetics, and localization precision. We screened 12 different DNA-PAINT sequences and used an optimized set to perform four-target 3D Exchange-PAINT in single E. coli cells. We further applied our strategy to image three effector mRNAs in the Gram-negative pathogen Yersinia enterocolitica and quantified RNA clusters and their 3D distribution. Our work provides a framework and practical guidelines for high-quality 3D DNA-PAINT imaging in Gram-negative bacteria and should facilitate its broader adoption in nanoscale bacteriology.

Our ability to efficiently construct synthetic DNA products is essential to studying and engineering biology. There is an abundance of widely adopted DNA assembly technologies suitable for generating synthetic constructs from diverse DNA inputs, but these techniques suffer from low specificity of connection during assembly, which decreases the accuracy, efficiency, and throughput of construction. Here we propose a broadly applicable paradigm for DNA assembly, High Temperature (HighT) assembly, that lowers misconnection rates by ligating complementary single strand DNA overhangs at reaction temperatures significantly higher than their predicted melting temperatures. By expanding the concept of melting temperature to demarcate smaller association fractions of DNA duplexes, we characterize HighT and demonstrate 40 percentage points increase in accuracy, and 4 orders of magnitude increase in yield of transformants when applying the technique to conventional assembly technologies. HighT is then applied to develop cloning pipelines for the development of efficient assembly-to-integration protocols, the construction of specific low-abundance isoforms of medically relevant proteins from human gDNA and oligos, and the de novo construction of a 10 kb, GC rich genomic region from the functionally extinct Northern White Rhinoceros created entirely in vitro.

Heterologous expression—the expression of genes from one organism into another—is an important approach to produce value-added chemicals such as proteins and metabolites. The main challenge, however, is achieving high production yields that are commercially viable. Approaches used to improve productivity rely on editing the DNA sequence, including promoters, ribosome binding sites, and codon usage. Epigenetics, in the form of DNA methylation, may offer an additional layer of control to improve production yields but remains unexplored. Moreover, the choice of the host organism is crucial. We have been interested in developing the alternative bacterial host Burkholderia sp. FERM BP-3421. Four DNA methyltransferases were predicted to be encoded in the genome of this strain, one of which was silent under our growth conditions. After performing gene deletions or overexpression, we experimentally assigned the recognition motif of each DNA methyltransferase, one of which was shown to be essential. As expected, modulation of DNA methylation had major impacts on the transcriptome and metabolome of the host, varying depending on which DNA methyltransferase was modified. Two of the methyltransferases that methylate adenine had overlapping profiles, whereas the one methylating cytosine differed. Finally, we probed the role of DNA methylation during heterologous expression using green fluorescence protein and three different promoters. Expression levels varied in a context-dependent manner with mechanisms, in some cases, involving the presence of methylation motifs within promoter regions but also based on global host effects. This work establishes epigenetics as an important design variable to help improve heterologous expression yields.

RNA splicing is frequently disrupted by genetic variation, contributing to a wide range of human diseases. Identifying the variants responsible for altered splicing is therefore crucial but has long been challenging. Deep learning models that predict splicing from genomic sequence have substantially advanced their detection, but accurately resolving their quantitative effects on splicing and the resulting transcripts remains an unmet need. Here we introduce SpliceAI2, a deep learning model trained end-to-end on splice site, splice junction, and transcript usage across multiple mammalian species. SpliceAI2 reconstructs complete transcripts directly from genomic sequence and considerably improves identification of splice-altering variants and quantification of their effects in diverse benchmarks spanning cryptic splicing, splicing quantitative trait loci, and massively parallel reporter assays. Human population datasets provide independent validation of these predicted functional effects through evidence of natural selection and associated changes in plasma protein abundance. Application to a rare disease cohort reveals that predicted splice-altering variants are highly enriched in disease-associated genes and account for 15% of the excess genetic burden, with approximately half residing in deep intronic regions likely missed by exome sequencing. Together, these findings establish SpliceAI2 as a major step forward in interpreting splice-altering variants across the human genome.

Conventional bio-diesel synthesis is both chemical- and energy-intensive, whereas lipase-catalyzed transesterification provides a more sustainable alternative. Nevertheless, the industrial application of native Candida antarctica lipase B (CalB) is limited by its low catalytic activity and high production costs. To overcome these challenges, we developed a recombinant Komagataella phaffii expression platform to efficiently secrete recombinant CalB (rCalB). To the best of our knowledge, this study demonstrates a proof-of-concept platform that integrates CalB secretion through a dual-promoter expression system regulated by AOX1/GAP in PDI-assisted and magnetically immobilized K. phaffii cells for biodiesel production. This integrated platform enabled extracellular production of rCalB and magnetic recovery of the engineered yeast cells. The maximum biodiesel yield was reached after 96 h using a methanol-to-oil ratio of 4:1 and a catalyst amount of 8 mL (2.4 EU) at 30 °C. The Gas Chromatography Flame Ionization Detector analysis results indicated that the produced biodiesel had higher amounts of C16 and C18 fatty acid methyl esters. These results highlight the potential to produce biodiesel via transesterification catalyzed by rCalB secreted from magnetically immobilized yeast cells.

Animals colonizing food-limited environments need to overcome growth constraints by adapting to low, inconsistent nutritional inputs. Comparative studies of surface and cave populations of the Mexican tetra fish species Astyanax mexicanus provide a powerful system to explore evolution to food scarcity. In nutrient-scarce caves, evolved traits such as eye loss and increased fat storage have been shown to conserve energy in cavefish. However, the role of microbiota-mediated metabolism in sustaining growth and survival under food limitation remains poorly understood. Here, we demonstrate that surface and cave populations distinctly modulate their gut microbiota, with cavefish recruiting microbes whose predicted metabolites specifically promote growth under nutrient-limited conditions. Using microbiota data from three distinct dietary regimes and a fasting treatment, integrated topological data analysis and sparse modeling identify sugar and fatty acid degradation pathways coupled with growth in cavefish. Notably, the metabolites of these pathways are beneficial only for cavefish growth, and not for surface fish. Fasting in cavefish led to the recruitment of gut microbes capable of metabolizing aromatic compounds inaccessible to the host, whereas it recruited few microbial associates in surface fish. Our findings indicate that cavefish evolved novel host-microbiota coupling to sustain growth and resilience in their nutrient-scarce niche. We anticipate that our findings will broadly inform studies of evolution under extreme nutritional conditions and provide insights into the role of host-microbiota interactions in resilience to food scarcity at varied time scales.

Metabolic cross-feeding is central to microbial life. Theory predicts that spatial structure favors cross-feeding by affording cells preferential access to the exchanged metabolites. Yet recent evidence shows that spatial structure can, in some contexts, disfavor cross-feeding. Here, we resolve this contradiction by structuring cross-feeding interactions into three dimensions: (i) directionality, (ii) investment, and (iii) metabolic dependency. We then investigated their combined effects using a spatially explicit individual-based model. We find that the effect of spatial structure on cross-feeding is determined not by physical proximity alone, but by how microbial growth depends on exchanged metabolites. Metabolic dependence shapes the spatial organization of cells and metabolite fluxes, and this emergent organization, in turn, reshapes the ecological outcome of cross-feeding. Our results reconcile previously conflicting observations and provide a general framework for predicting and engineering metabolically interdependent microbial communities.

Programmable RNA sensors enable sequence-specific detection of RNA inputs for diagnostics, genetic circuit control, and multiplexed molecular readouts. Although readily programmable, their performance varies substantially with the target site selected within a transcript, making target-site selection a major challenge. Here, we report DeSTAR (Detectron Sensor TARget-site ranker), a machine-learning-assisted target-site ranking model for programmable RNA sensors. To train this model, we generated transcript-wide, single-nucleotide-resolution tiling sensor libraries across six diverse RNA targets, enabling generalization to be tested across transcripts rather than within a single RNA. In leave-one-transcript-out evaluation, DeSTAR's top-ranked sites were enriched for high-performing sensors in held-out transcripts, driven predominantly by interpretable sequence grammar in the sensor stem rather than transcript-specific patterns. Prospectively, on an unseen SARS-CoV-2 transcript, the model's top-ranked sites yielded working sensors roughly five-fold more often than random selection. These results support DeSTAR as a generalizable strategy for target-site selection in RNA sensors across transcripts.

tRNA are short, ~76 nucleotide-long, modification-rich molecules that play an important role in protein translation. Nanopore sequencing is the only platform capable of directly sequencing full-length, native tRNA molecules. This capability is enabled by splinted ligation of tRNA-specific adapters. As a result, the measured ionic current signal reflects a composition of both adapter and tRNA sequence features. Despite these advantages, analysis of nanopore tRNA sequencing data poses distinct computational challenges. Existing ionic current-based signal-to-sequence decoding (basecalling) models are optimized for mRNA and perform suboptimally on short, modification-dense RNA such as tRNA. Furthermore, alignment algorithms for reference-guided analysis of short-length reads are not optimized for nanopore reads with lower accuracy than Next-generation sequencing reads. To overcome these limitations, domain-specific computational approaches are required. We present tRNAZAP, a nanopore ionic current signal analysis toolkit designed to enable precise reference-guided sequence alignment. tRNAZAP operates directly on raw ionic current signals and decomposes the analysis into two complementary tasks: 1) segmentation of adapter and tRNA regions, and 2) tRNA classification, both implemented using transformer-based models. The segmentation component isolates the tRNA signal portion and extracts the corresponding nucleotide sequence as the alignment query, while the classification component identifies the tRNA type as the alignment reference without requiring sequence-level information. Together, these outputs allow direct optimal pairwise alignment between the tRNA subsequence of the full query and reference sequences. Trained on in vitro transcribed RNA data and evaluated on yeast tRNA, these models achieved ~98% classification accuracy and intersection-over-union of ~0.98 for segmentation.

Interactions between microbes shape the structure and function of microbial communities. While studying interactions is key to understanding microbial communities as a whole, gaining a detailed mechanistic view is challenging due to the number of co-occurring interactions. Here, we focus on a single class of shared nutrients, the cobalamin (vitamin B12) family of enzyme cofactors known as cobamides, to study nutrient interactions in laboratory consortia of increasing complexity. Measuring cobamide-dependent growth of two soil Caulobacter isolates in monoculture revealed that each grows faster within a distinct range of cobamide concentrations. Co-culturing the two isolates demonstrated that the cobamide concentration predictably determines which isolate predominates. These results suggest that, although the two organisms appear functionally redundant, with identical metabolic capabilities predicted in their genomes, they may occupy distinct niches in their environment based on cobamide concentration. We further examined the ability of cobamide-producing Mesorhizobium and Priestia isolates to share cobamides and found that both support the cobamide-dependent growth of the two Caulobacter isolates. The Caulobacter isolates could coexist in tri-cultures with each producer following serial passaging, indicating the producers can support the growth of multiple cobamide-dependent organisms. We analyzed the metabolic capacity encoded in the genomes of the four isolates and found that cobamides are likely the main shared nutrient in our co- and tri-cultures. These results highlight the utility of cobamides as a model nutrient to characterize and predict interactions in bacterial consortia of increasing complexity.

Metagenomic databases hold vast quantities of unexplored genetic material that can seed tools for biotechnology and medicine, but discovery is limited by our ability to examine sequence data at scale. Here, we show that large language models (LLMs) in an autonomous harness can alleviate this bottleneck and surface novel molecular systems. We deployed Claude Code instances to survey reverse transcriptase (RT) loci across 1.9 billion protein clusters. Among the top candidates were array-associated RTs (ART), a novel family of jumbo-phage RTs coupled with an array of ~200-nucleotide units and a dedicated partner gene. Session transcripts show that Claude identified the ART family by directly examining its DNA sequences and recognizing the atypical repeats, a behavior attributable to specific Mythos 5 internal signals that respond to repeated DNA. ART arrays are highly expressed and appear as discrete units during Staphylococcus phage infection, suggesting an RT system directed by a repertoire of distinct RNAs. These findings indicate that LLMs can autonomously detect anomalies and drive analyses to initiate biological discoveries.

Protein structural complexity confers responsiveness and modularity, allowing proteins to act as control modules in biomolecular circuitry. Protein engineering that commandeers native cell signaling and central dogma functions has led to the development of cell therapies, specialized biosensors, and a deeper understanding of cellular mechanisms. In this review, we examine how protein control modules command circuit with therapeutic functionality across cellular spatial levels. Extracellularly, signal proteins such as antibodies and cytokines can provide either binary or gradient-sensitive detection of environmental features. Extracellular proteins also serve as inputs to synthetic receptors that transmit signals across membranes, rapidly linking environmental cues to intracellular action. Inside the cell, protein control modules are engineered from central dogma components to bind nucleic acids, creating gene circuits that influence transcription, translation, and nucleic acid processing, and ultimately regulate protein expression. This review covers circuitry platforms and immunoengineering applications across spatial levels, focusing on innovation from 2016 to 2026. With continued development, synthetic biocircuits can serve as a platform for creating next-generation “smart” therapeutics and ultimately for engineering immunity.

Protein inverse folding, which infers compatible amino acid sequences from a given backbone structure, is central to rational protein design. Current methods rely largely on sequence recovery rate as the evaluation metric, emphasizing structural compatibility but seldom addressing function. Yet functional activity depends on the proper network of non-covalent interactions-protein-ligand contacts being a prime example. We argue that function-oriented protein design is governed by three complementary constraints: structural, interaction, and evolutionary, none of which are handled in a unified manner by existing pretrained models. To bridge this gap, we introduce zPolaron, a graph neural network built on protein-ligand complexes for functional protein design. zPolaron represents residues and ligand atoms as a heterogeneous graph and jointly encodes structural, interaction, and evolutionary information by fusing physical-interaction-enhanced graph representations with ESM-2 embeddings. Across multiple fitness prediction benchmarks, zPolaron outperforms all baselines, with gains evident across structural, interaction, and evolutionary dimensions. Wet-lab experiments further confirm its utility in functional enzyme screening and design: the optimized CalA6-D205A-S160G mutant achieves a 17.84-fold improvement in activity, while the CalB-I121M mutant yields a 1.91-fold increase. Together, these results show that integrating interaction and evolutionary information fills the missing puzzle of inverse folding, enhancing protein fitness prediction beyond structural compatibility. zPolaron is available at https://github.com/zelixirSH/zPolaron.