The core problem the researchers confronted is a familiar one in genomics: experimental annotation is slow and expensive, while genome sequencing is fast and cheap. When a new plant genome is sequenced, thousands of proteins are catalogued with little more than a name and a sequence. Computational tools can transfer functional labels from well-studied relatives, but this homology-based approach breaks down for proteins that have no close characterized counterparts, which is precisely the situation in many orphan crops. General-purpose deep learning function predictors, trained on multi-taxon datasets, tend to underrepresent experimentally annotated plant proteins, so their celebrated performance on animals and microbes does not necessarily transfer to the green lineage. DeepGreenGO was built to close that transferability gap by learning the patterns of plant protein function in a dataset where plants are not a footnote but the entire curriculum.

Technically, the model combines two complementary views of a protein. The first comes from ProtBERT-BFD, a protein language model trained on the enormous Big Fantastic Database, which converts each amino acid residue into a rich numerical embedding that encodes evolutionary and biochemical context learned from billions of sequences. The second view is structural: the researchers derived contact maps from protein structures, essentially matrices recording which residues lie close together in three-dimensional space. These contact maps define a graph in which residues are nodes and physical proximity defines the edges. This graph representation is then processed by two types of graph neural network layers in sequence: graph convolutional network layers, which aggregate information from neighboring residues, and GATv2 layers, a more expressive form of graph attention that learns to weight which neighbors matter most for each residue. An attention pooling step then compresses the residue-level representations into a single protein-level vector that feeds a multilabel classifier.

The output of that classifier is a set of Gene Ontology terms, the standardized vocabulary that biologists use to describe protein functions across three branches: molecular function, biological process, and cellular component. Because Gene Ontology is hierarchical, predicting a function is inherently a multilabel problem, and DeepGreenGO was trained to predict terms specific to the ontology it was trained on. The training corpus was a carefully curated Viridiplantae dataset of 7,534 experimentally annotated protein structures drawn from the Protein Data Bank, cross-referenced through the SIFTS resource to link structures to sequences and taxonomy. Crucially, the team split these proteins into training, validation, and test sets of 6,026, 754, and 754 PDB chains respectively using sequence-similarity-aware clustering, a methodological safeguard that prevents near-identical proteins from appearing in both training and test data and inflating performance estimates.

The benchmarking results reveal where the plant-specific approach pays off. DeepGreenGO was evaluated against sequence homology-based tools and against other recent sequence- and structure-based deep learning methods, and it showed its strongest performance in the biological process ontology, arguably the hardest and most informative of the three GO branches. There it achieved the highest protein-centric maximum F1 score, known as Fmax, of 0.227, and the lowest Smin score of 19.73, a metric that penalizes both missed and spurious predictions in a semantically aware way. Perhaps most striking was its showing on the information-content weighted area under the precision-recall curve, where it scored highest among all compared methods for biological process. That metric rewards models for correctly predicting rare, informative annotations rather than playing it safe with broad, shallow terms, suggesting that DeepGreenGO is genuinely learning specific plant biology rather than recycling generic predictions.

Ablation experiments, in which components of the model are systematically removed to measure their contribution, added an honest and instructive note to the study. The pretrained ProtBERT-BFD sequence embeddings turned out to provide most of the predictive signal, confirming the now well-established principle that self-supervised protein language models capture a remarkable amount of functional information on their own. The structural contact maps, processed through the graph neural network stack, contributed more modestly, and the authors identify improving the integration of structural information as a clear opportunity for future work. This kind of transparency matters in a field where architectural novelty is often claimed as the source of gains; here the data show that the language model backbone is central, while the graph machinery offers a framework that can be sharpened as structural prediction quality improves across the plant kingdom.

To demonstrate that the model is more than a benchmark winner, the researchers deployed DeepGreenGO across the entire proteome of rice, Oryza sativa, screening all 43,649 rice proteins for functions related to seed development. The scan flagged 372 candidate proteins associated with this agriculturally critical process. The team then validated the biological plausibility of these predictions through two independent lines of evidence. Functional enrichment analysis showed that the predicted proteins were collectively linked to biological processes known to operate during seed formation, and transcriptomic analysis revealed that many of them display elevated expression in ovary, embryo, and endosperm tissues, the anatomical arenas where seed development unfolds. For a purely computational prediction to align with tissue-level expression patterns is a meaningful consistency check, and it suggests the candidate list is a credible starting point for experimental follow-up.

The implications reach well beyond rice. Seed development genes influence yield, grain quality, and stress resilience, traits at the heart of breeding programs aimed at sustainable agriculture in a changing climate. A tool that can prioritize which of tens of thousands of uncharacterized proteins deserve scarce laboratory resources is a practical accelerator for crop science, particularly for orphan crops and underexplored species where homology-based annotation fails. Because DeepGreenGO is taxonomically focused, its predictions are also more trustworthy for proteins with limited detectable similarity to the training set, exactly the proteins that dominate non-model plant genomes. The framework could in principle be retrained or extended as new plant structures are experimentally solved and as AlphaFold-style predicted structures expand the structural universe available to graph-based methods.

There are also broader lessons here for the design of biological AI systems. The study is a case study in domain-specific curation: rather than scraping the largest possible dataset, the team built a plant-only corpus, controlled for sequence similarity during data splitting, and evaluated with metrics that distinguish informative predictions from easy ones. The modest absolute Fmax values in the biological process ontology are a candid reminder that protein function prediction remains genuinely hard, and that headline numbers from multi-taxon benchmarks should not be assumed to hold in specialized domains. The work also highlights the growing role of research groups outside the traditional centers of AI and genomics; the Sri Lankan team, working with donated workstations acknowledged from the Colombo University Faculty of Science Alumni Association in North America and a faculty high-performance computing facility, produced a competitive model without industrial-scale resources.

DeepGreenGO arrives at a moment when the intersection of protein language models, graph learning, and structural biology is reshaping what is computable in the life sciences. Its strongest advantages, in informative biological process annotations and in proteins far from any characterized relative, point toward a future where plant genomes are no longer annotated by proxy from animals and fungi but on their own terms. The rice seed development results offer a template for how such models should be deployed: predict broadly, then triangulate with enrichment and expression data before anyone commits a greenhouse season to a candidate gene. As plant structural datasets grow and structural integration improves, models of this kind could become standard instruments in the crop biologist’s toolkit, quietly converting sequence catalogs into testable biology, one green protein at a time.

Subject of Research: Graph neural network-based deep learning for plant-specific protein function prediction

Article Title: DeepGreenGO: a graph neural network-based deep learning model for plant-specific protein function prediction

Article References: Sridharan, G., Weththasinghe, S. A., Sridharan, A., Upeka, W. M. M., Abeywardhana, D. L., & Fernando, P. C. (2026). DeepGreenGO: a graph neural network-based deep learning model for plant-specific protein function prediction. BMC Bioinformatics. https://doi.org/10.1186/s12859-026-06677-9

Image Credits: AI Generated

DOI: 10.1186/s12859-026-06677-9

Keywords: DeepGreenGO, protein function prediction, graph neural networks, plant biology, Gene Ontology, ProtBERT-BFD, rice, seed development, bioinformatics, deep learning, Viridiplantae, sustainable agriculture

Tem1 is a small GTPase, a class of molecular switches that alternate between an active GTP-bound state and an inactive GDP-bound state. In budding yeast, Tem1 sits at the top of the mitotic exit network, or MEN, a signaling cascade that couples the completion of nuclear division to cytokinesis. The Bfa1-Bub2 complex acts as a GTPase-activating protein, or GAP, that hydrolyzes Tem1-bound GTP, keeping the pathway off until the mitotic spindle is correctly positioned. When the spindle pole body, the fungal equivalent of the centrosome, reaches the right location, the polo-like kinase Cdc5 disables the GAP, allowing GTP-loaded Tem1 to trigger a kinase cascade that ultimately releases the phosphatase Cdc14 and drives the cell out of mitosis. Homologs of this pathway exist across fungi, but until now M. oryzae lacked a confirmed Tem1 ortholog, leaving a critical gap in understanding how the pathogen coordinates its unusual cell biology with infection.

To find the missing piece, the team searched the M. oryzae genome using yeast and Fusarium graminearum Tem1 proteins as queries. The search pinpointed MGG_04862, annotated as a septum-promoting GTP-binding protein, which shares 60.21 percent sequence identity with yeast Tem1 and a striking 72.28 percent identity with the Fusarium homolog. Phylogenetic analysis of Tem1 proteins from 24 fungal species placed MoTem1 in a clade with homologs from other devastating plant pathogens, including Botrytis cinerea and Sclerotinia sclerotiorum. The researchers then built a panel of isogenic strains in the Guy11 wild-type background: a knockout lacking the gene entirely, an overexpression strain, and two point mutants locked into opposite states. The Q182L substitution produces a constitutively active GTP-bound protein, while T137N creates a dominant-negative, inactive version. Yeast two-hybrid and pull-down assays confirmed that MoTem1 interacts with the fungal GAP components MoBfa1 and MoBub2 in a state-dependent manner, with the active form recruiting the full complex and the inactive form binding only MoBfa1.

The phenotypes of these mutants revealed that MoTem1’s activity cycling, rather than its mere presence or abundance, is what matters. The constitutively active strain grew about 41.56 percent more slowly than wild type, while the knockout and both locked-state mutants produced dramatically fewer conidia, the three-celled spores that launch infections, with reductions of 35 to 67 percent. Overexpression alone left conidiation untouched, demonstrating that dynamic GDP/GTP cycling is essential for asexual reproduction. Germination assays told a similar story: strains with enhanced or reduced GTPase activity germinated faster than wild type, reaching up to nearly 90 percent germination within two hours, while the constitutively active strain germinated poorly, formed multiple germ tubes, developed thickened cell walls and produced abnormally elongated germ tubes. These defects in pre-infection morphogenesis foreshadowed even more dramatic consequences on the plant itself.

When the researchers spray-inoculated rice seedlings, the virulence patterns were striking. The knockout and constitutively active strains were both attenuated, with the active mutant restricted to small lesion types and never producing the largest class of lesions. The dominant-negative strain, however, was hypervirulent, generating large Type IV lesions that were absent in wild-type infections at the same time point. The authors propose that inactivating MoTem1 promotes invasive hyphal growth, the phase of infection during which the fungus spreads from cell to cell through plasmodesmata, and lesion expansion directly depends on how extensively those hyphae extend. Constitutive activation, by contrast, disrupts the timing of morphogenic programs and likely the deployment of effectors, undermining the finely choreographed sequence of appressorium formation, cuticle penetration and biotrophic establishment that M. oryzae requires.

Stress experiments added another layer of complexity. Under oxidative stress from hydrogen peroxide, the dominant-negative strain was significantly more tolerant than wild type, with a higher half-maximal inhibitory concentration, while the constitutively active strain was hypersensitive. Cell wall stress from Congo Red produced the same split: overexpression conferred tolerance, constitutive activation caused severe sensitivity. Under osmotic stress with sorbitol, the knockout, overexpression and dominant-negative strains all showed enhanced tolerance, but responses to ionic stress with sodium and potassium chloride varied by allele. The picture that emerges is of a cell cycle regulator whose nucleotide-binding state feeds into entirely separate stress signaling pathways, with the GTP-bound form promoting oxidative resilience and the cycling state maintaining cell wall and membrane integrity.

Fluorescence microscopy explained part of the mechanism. EGFP-tagged MoTem1 localized as puncta to spindle pole bodies, confirmed by co-localization with the SPB marker MoAlp6-mCherry. The constitutively active variant accumulated intensely at these structures in both hyphae and conidia, while the inactive variant showed completely diffuse cytoplasmic fluorescence, demonstrating that only the GTP-bound form is recruited to spindle pole bodies. Biochemical assays confirmed that purified MoTem1 has genuine GTP-hydrolyzing activity, roughly four times that of a tag-only control, and that the T137N substitution abolishes it. Critically, both the knockout and the constitutively active strain produced multinucleate hyphal compartments, indicating that nuclear division and septum formation had fallen out of sync. In the knockout, cytokinesis fails outright; in the active mutant, premature MEN activation bypasses the spindle position checkpoint, or SPOC, the surveillance system that normally delays mitotic exit until the spindle is correctly aligned. The dominant-negative strain maintained normal nuclear segregation, which the authors attribute to residual wild-type MoTem1 activity in that background being sufficient for cytokinesis.

Transcriptomic profiling across all five strains revealed the full scope of MoTem1’s influence. Principal component analysis separated the knockout and constitutively active strains sharply from wild type, and differential expression analysis identified 72 genes consistently altered across all mutants, including core cell cycle regulators and stress-responsive transcription factors. Weighted gene co-expression network analysis resolved eight modules, three of which stood out: a brown module correlated with oxidative stress resistance and negatively with growth rate, suggesting a trade-off between stress adaptation and proliferation; a yellow module tied to hyphal growth and enriched for ribosome and glycolysis genes; and a tan module most strongly correlated with pathogenicity, containing 38 virulence-associated genes encoding secreted effectors and cutinases. Among the tan module’s hub genes was MoCHS1, a class V chitin synthase, one of three of M. oryzae’s seven chitin synthases known to be required for pathogenicity.

The link to MoCHS1 proved to be the study’s most consequential finding. Expression of this chitin synthase was downregulated 3.07-fold in the constitutively active strain and upregulated 1.26-fold in the dominant-negative strain, mirroring their virulence phenotypes. Pharmacological validation followed: the chitin synthase inhibitor Polyoxin B suppressed growth of the constitutively active strain by over 40 percent while barely touching wild type, and the dominant-negative strain was strongly resistant. Reactive oxygen species assays on rice leaf disks added a plant immunity dimension. Culture filtrates from the dominant-negative strain elicited the strongest ROS burst in rice tissue, followed by wild type, with the constitutively active strain eliciting the weakest response, and chitinase pretreatment abolished all responses, confirming that chitin oligomers acting as pathogen-associated molecular patterns were responsible. The authors suggest that altered MoCHS1 expression changes how much chitin is exposed or released from the fungal cell wall, modulating the plant’s early immune perception and potentially accelerating the transition to the necrotrophic phase that produces large necrotic lesions.

Taken together, the study repositions MoTem1 from a narrow cell cycle component to a signaling hub with genome-wide reach. Its GTP-dependent recruitment to spindle pole bodies safeguards mitotic fidelity, while its activity states ripple outward through transcriptional networks governing redox balance, cell wall architecture, osmotic adaptation and virulence. The work also highlights evolutionary divergence within the MEN pathway: unlike in Fusarium graminearum, where inactive Tem1 localizes to septa, inactive MoTem1 disperses into the cytoplasm, and M. oryzae’s MoSep1 activates MoMob1 directly, bypassing MoDbf2. For a pathogen responsible for one of agriculture’s most serious diseases, these findings open a conceptually new angle: the machinery that ends mitosis is also a master regulator of how the fungus grows, withstands stress and attacks its host, and perturbing its nucleotide cycle can either cripple the pathogen or, paradoxically, make it more dangerous.

Subject of Research: Role of the mitotic exit GTPase MoTem1 in development, stress adaptation and pathogenicity of the rice blast fungus Magnaporthe oryzae

Article Title: Mutational analysis of the mitotic exit GTPase MoTem1 reveals its role in development, stress adaptation, pathogenicity and global gene regulation in Magnaporthe oryzae

Article References: Pei, M., Xie, X., Cao, Y., Chen, J., Yang, F., Wang, Z., Olsson, S., Lu, G.-D., & Li, Y. (2026). Mutational analysis of the mitotic exit GTPase MoTem1 reveals its role in development, stress adaptation, pathogenicity and global gene regulation in Magnaporthe oryzae. Stress Biology, 6(1), Article 38. https://doi.org/10.1007/s44154-026-00310-8

Image Credits: AI Generated

DOI: 10.1007/s44154-026-00310-8

Keywords: Magnaporthe oryzae, MoTem1, mitotic exit network, GTPase, spindle pole body, SPOC, chitin synthase, MoCHS1, pathogenicity, stress adaptation, transcriptomics, WGCNA

The limitations of conventional breeding are stark. Hybridization-based pyramiding of favorable genes typically requires ten to fifteen years per cycle, and screening efficiency is low. Worse, genetic linkage often ties desirable genes to undesirable neighbors on the same chromosome, producing genetic drag that breeders cannot easily escape. Because recombination during crossing is random, traditional methods cannot target multi-gene networks or create rare combinations of favorable alleles that simply do not exist in nature. Crop yield itself is not governed by a single gene but emerges from synergistic interactions among growth, disease resistance, pest resistance and stress tolerance networks, while quality traits involve nutrition, taste and appearance simultaneously. Single-trait improvement, the review concludes, leaves an inherent gap between what breeding can deliver and what modern agriculture needs.

Multiplex genome editing closes that gap by exploiting the core mechanics of the CRISPR/Cas system. A single guide RNA forms a ribonucleoprotein complex with a Cas endonuclease, recognizes target double-stranded DNA through base pairing and a protospacer adjacent motif, and induces a site-specific double-strand break. Cellular repair then proceeds through error-prone non-homologous end joining, whose small insertions and deletions can knock out a gene by frameshifting its reading frame, or through template-directed homology-directed repair. The decisive advantage of MGE is that a single vector, delivered in one transformation, can carry guides for many loci at once, coordinating expression across pathways without touching unrelated genes.

The technology’s power is clearest in cases where single-gene edits fail. Genetic redundancy means that mutating one member of a gene family often produces no visible phenotype because homologous genes compensate. In wheat, broad-spectrum resistance to powdery mildew was achieved only by simultaneously mutating all three homoeoalleles of the TaMLO gene family across the A, B and D genomes; plants with a single copy knocked out remained susceptible. The same principle holds in Arabidopsis, where the AtMLO2/6/12 triple mutant shows markedly enhanced pathogen resistance. The review traces the field’s milestones from early two-guide experiments in Arabidopsis and rice, through the tRNA-sgRNA tandem arrays that let one transcript yield many mature guides, to the current record of thirteen target sites edited simultaneously in rice.

Three co-expression architectures dominate the toolkit. The Multi-Transcriptional Unit system drives each sgRNA with its own RNA polymerase III promoter, offering modular control but risking plasmid recombination from repeated promoter sequences and producing bulky vectors. The Twin-Transcriptional Unit system separates Cas9 and sgRNA expression, processing a single polycistronic precursor with tRNA, Csy4 ribozyme or Hammerhead elements; the tRNA strategy, which hijacks the plant’s own RNase P and RNase Z machinery, achieved over fifty percent simultaneous mutation of three genes in rice and supports a modular toolkit for up to eight targets. The Single Transcriptional Unit system fuses Cas9 and sgRNAs into one compact transcript, ideal for species like soybean that tolerate vector redundancy poorly. Species choice matters: monocots such as rice and wheat favor MCTU and TCTU designs, while dicots often require species-specific Pol II promoters and careful control of tandem target numbers.

Applied outcomes span every major breeding goal. In disease resistance, simultaneous targeting of the Bsr-d1, Pi21 and ERF922 genes produced rice varieties resistant to both blast and bacterial blight, with field durability two to three years longer than single-gene lines, because pathogens must overcome multiple infection barriers at once. In abiotic stress, editing three genes governing root development, grain shape and cold response yielded rice that lost only 3.2 percent of yield under cold stress, compared with fifteen to twenty percent losses in conventional cold-tolerant varieties. Tomato plants edited at the HyPRP1 gene gained over forty percent salt tolerance plus cross-resistance to drought and cold. Yield work has been equally striking: editing GS3, TGW3 and GW8 produced high-yielding hybrid rice with slender grains, while combining eight sgRNAs against cytochrome P450 homologs and the BADH2 gene boosted yield and endowed grains with a fragrant aroma.

Quality improvement may be the most commercially compelling frontier. In soybean, knocking out the GmFAD2-1A and GmFAD2-1B desaturase genes created high-oleic-acid oil with better oxidative stability, and stacking those edits with lipoxygenase knockouts removed the beany flavor that limits soy products. Combining two multiplex-edited soybean varieties allowed researchers to tune protein functionality, including emulsifying activity, gelation and solubility, by editing storage protein genes such as Glycinin and beta-Conglycinin. In tomato, simultaneous editing of the PSY1, MYB12 and SGR1 genes recolored fruit and improved nutrition, while triple knockout of LCY, SGR1 and BLC1 blocked carotenoid branch metabolism and raised lycopene accumulation fivefold. Wheat edited across eight gliadin genes points toward hypoimmunogenic bread, and knocking out three soybean lipoxygenase genes eliminated off-flavors.

Perhaps the most visionary application is de novo domestication. Because centuries of artificial selection narrowed the genetic base of crops and discarded valuable stress-resistance traits found in wild relatives, MGE offers a shortcut: edit domestication genes directly in wild germplasm. Researchers established the first de novo domestication system for wild allotetraploid rice, Oryza alta, editing homologs controlling seed shattering, awn length, plant height and grain traits. Even more dramatic, the salt-tolerant landrace Sea Rice 86 was improved by synchronously editing thirteen key agronomic genes in a single transformation; the resulting homozygous line retained its salt tolerance while gaining improvements in plant height, architecture, grain shape, yield components and photoperiod sensitivity. Wild tomato has likewise been rapidly domesticated with eight sgRNAs targeting six agronomic genes, producing larger fruits and higher lycopene content.

Challenges remain substantial. Gene interactions can be antagonistic as well as synergistic: editing rice grain length, width and number genes together raised yield, but loss of Gn1a overactivated cytokinin signaling, multiplying tillers and undermining lodging resistance, a problem that had to be solved by recombining targets such as Ghd7, DTH8 and OsNAC006. Editing efficiency decays with scale, holding at fifty to seventy percent for three to five genes but dropping below thirty percent beyond eight targets. New tools are responding: Cas12a uses short crRNAs that cut sequence redundancy by more than half, processes its own arrays without exogenous enzymes, and engineered variants like Mb3Cas12a now work across relaxed PAM sites and low temperatures. Prime editing, which installs precise changes without double-strand breaks, achieved 28.6 percent dual-site and 7.1 percent quadruple-site efficiency in rice, while an ultra-efficient tomato system reached 87.5 percent single-target precision with multiplex rates matching single edits.

Delivery and design are the next battlegrounds. Agrobacterium transformation and particle bombardment remain mainstream but struggle with large multiplex vectors and recalcitrant species, and both depend on laborious tissue culture. Virus-induced genome editing offers a way out: a Barley stripe mosaic virus system delivering sgRNAs into Cas-expressing wheat reached up to 47.3 percent dual-gene editing in the first generation, with over eighty percent of mutants virus-free by the next, while SYNV vectors can carry the entire CRISPR/Cas machinery in one round. On the design side, machine learning and large language models such as CRISPR-GPT are automating sgRNA selection, and AlphaFold-guided editing has already produced high-oil soybean alleles by predicting which mutations preserve protein structure. The review’s authors frame the trajectory clearly: as efficiency bottlenecks fall and intelligent design matures, multiplex genome editing is propelling crop breeding from experience-dependent craft toward precision engineering of complex, customized traits, a shift with consequences for global food security that are only beginning to unfold.

Subject of Research: CRISPR/Cas-mediated multiplex genome editing for pyramiding multiple agronomic traits in crop breeding

Article Title: Multiplex gene editing drives revolution in crop breeding: overlaid editing of multiple genes and customization of complex traits

Article References: Multiplex gene editing drives revolution in crop breeding: overlaid editing of multiple genes and customization of complex traits. (n.d.). https://doi.org/10.1007/s44307-026-00099-7

Image Credits: AI Generated

DOI: 10.1007/s44307-026-00099-7

Keywords: CRISPR/Cas9, multiplex genome editing, crop breeding, trait pyramiding, Cas12a, prime editing, de novo domestication, rice, soybean, tomato, wheat, virus-induced genome editing

ASFV belongs to the nucleocytoplasmic large DNA viruses, a group of complex viruses whose genomes encode many of their own replication machinery. Although the virus assembles and replicates in cytoplasmic factories, its entry into a macrophage sets off a cascade of events that reaches into the nucleus. Previous work had suggested that ASFV can modulate host gene expression epigenetically, but the scale and dynamics of the nuclear changes during early infection had never been systematically mapped. The new study addresses that gap by tracking the infection at 2, 4, and 6 hours post-infection, a window in which the virus establishes itself before progeny virions begin to spread.

To capture this pivotal phase, the researchers deployed an unusually comprehensive multi-omics strategy. Time-course RNA sequencing charted the host transcriptome as it changed hour by hour. CUT&Tag profiling measured where RNA polymerase II, the enzyme that transcribes protein-coding genes, was sitting across the genome. Immunoprecipitation followed by mass spectrometry identified the proteins that associate with the polymerase during infection. Finally, Hi-C, a method that captures the physical contacts between distant stretches of chromatin, revealed how the genome folds in three-dimensional space. Crucially, the team ran this entire pipeline using both a virulent ASFV strain and its live-attenuated derivative, allowing them to distinguish changes caused by the virus in general from those attributable to specific virulence factors.

The most striking finding concerns the global organization of the genome. In healthy cells, chromatin is partitioned into compartments of active euchromatin and silent heterochromatin, and further subdivided into topologically associating domains, or TADs, which are self-contacting genomic neighborhoods that help regulate which genes talk to which regulatory elements. The study found that early ASFV infection significantly increased heterochromatinization across the host genome, shifted the positions of TAD boundaries, and altered both the intensity of chromatin interactions and the spatial distances between interacting genomic regions. In other words, within hours of the virus entering the cell, the host genome was physically refolded.

One of the most intriguing observations is that viral DNA itself appears to participate in this reorganization. The Hi-C data revealed interactions between viral genomes and host chromatin, suggesting that incoming ASFV DNA does not simply float in the nucleus as an inert passenger. Instead, it seems to make physical contact with host chromosomes, raising the possibility that the virus actively recruits host genomic regions to viral replication compartments or sequesters immune genes into repressive nuclear environments. The authors suggest that this chromatin reorganization may contribute to ASFV immune evasion, though the precise mechanisms linking physical genome folding to immune suppression remain to be worked out.

The transcriptional consequences of this remodeling were equally revealing. The host antiviral transcriptional program, the set of interferon-stimulated genes and inflammatory pathways that normally swing into action when a macrophage detects a pathogen, turned out to be indistinguishable between cells infected with the virulent strain and cells infected with the attenuated virus. This was a surprise, because the attenuated strain lacks MGF-encoded factors, a family of viral genes long associated with virulence and immune modulation. The result demonstrates that the earliest wave of innate immune activation is independent of these virulence determinants, implying that whatever the MGF proteins do to tip the balance toward lethal disease must happen later, or through mechanisms other than silencing the initial antiviral transcriptional response.

RNA polymerase II itself became a focal point of the analysis. The researchers found that the total protein levels of the polymerase did not change appreciably during early infection, yet its behavior was profoundly altered. Its genomic occupancy was redistributed across the genome, and its interactome, the network of proteins it physically associates with, was remodeled. Core polymerase subunits and components of the Mediator complex, a large coactivator that bridges transcription factors to the polymerase, became enriched in the polymerase-associated fraction. In contrast, sequence-specific transcription factors, the proteins that normally direct the polymerase to particular genes in response to signals, were depleted from that fraction.

This shift in the polymerase’s social circle is mechanistically significant. A polymerase that is increasingly bound to Mediator and core subunits but less associated with gene-specific transcription factors suggests a reprogramming of transcriptional control away from signal-responsive gene activation and toward a more basal or virally steered mode. Combined with the increased heterochromatinization and the repositioning of TAD boundaries, the picture that emerges is one of coordinated change: the physical folding of the genome, the distribution of the transcription machinery, and the resulting transcriptome are all being reshaped together during the first six hours of infection.

The implications extend beyond basic virology. African swine fever continues to inflict substantial economic losses on the global swine industry, and the absence of a widely available safe and effective vaccine has made understanding the virus’s evasion strategies a research priority. By showing that ASFV manipulates host chromatin architecture so early and so extensively, the study opens a new dimension for antiviral intervention. If specific host factors mediate the observed heterochromatinization or the redistribution of polymerase occupancy, those factors could become drug targets. Similarly, the finding that early innate immune activation proceeds normally even without MGF-encoded virulence factors suggests that the critical battle between host defense and viral evasion is fought on a chromatin landscape, not merely at the level of individual signaling pathways.

The work also adds ASFV to a growing list of viruses known to touch the three-dimensional genome, but with a distinctive twist. Because ASFV is a cytoplasmic replicator, its nuclear influence must be exerted through delivered proteins, RNAs, or the viral DNA itself, rather than through a nuclear replication cycle. Disentangling which viral components drive the chromatin remodeling, and determining whether blocking that remodeling restores full antiviral gene expression, are the natural next questions. For now, the study provides the most systematic view yet of how a lethal agricultural pathogen rewrites the physical grammar of the host genome in its opening hours, and it suggests that the nucleus, long viewed as a bystander in ASFV infection, is in fact an active battlefield.

Subject of Research: Chromatin architecture remodeling in host cells during early African swine fever virus infection

Article Title: ASFV early infection dynamically remodels host chromatin architecture to evade immune responses

Article References: Xu, Y., Yang, B., Zhu, J., Sunkang, Y., Xing, L., Wang, C., Yang, J., Xu, W., Dai, J., Zhang, J., Cao, G., & Xiao, K. (2026). ASFV early infection dynamically remodels host chromatin architecture to evade immune responses. Cellular and Molecular Life Sciences. https://doi.org/10.1007/s00018-026-06467-9

Image Credits: AI Generated

DOI: 10.1007/s00018-026-06467-9

Keywords: African swine fever virus, ASFV, chromatin architecture, Hi-C, RNA polymerase II, macrophages, immune evasion, topologically associating domains, heterochromatin, innate immunity, virus-host interactions, multi-omics

Nuruk is the cornerstone of traditional Korean alcohol production. This naturally fermented starter cake, typically made from wheat or rice and allowed to colonize with a diverse community of molds, yeasts, and bacteria, supplies the enzymes that break down starches in grains into fermentable sugars. Because nuruk is produced through spontaneous microbial growth, mashes made with it contain a rich and variable cocktail of microorganisms, which in turn generates a wide spectrum of flavor precursors. When such a mash is distilled, the volatile compounds formed during fermentation, including esters, higher alcohols, organic acids, and aldehydes, are carried into the distillate and define the spirit’s aroma. This makes the fermentation stage, and especially the yeast population within it, a decisive lever for controlling the sensory quality of the final product.

In the new study, Son set out to test whether commercially available wild yeasts could act as that lever in a controlled and reproducible way. The experimental design compared three commercial non-Saccharomyces yeasts, Lachancea thermotolerans, Metschnikowia pulcherrima, and Torulaspora delbrueckii, with a mixed-yeast preparation combining Saccharomyces cerevisiae and Torulaspora delbrueckii, and with a standard brewing yeast. A control fermentation relying on the native microbes of the nuruk mash completed the set. After fermentation, all of the mashes were processed under identical conditions using vacuum distillation, a technique that lowers the boiling point of ethanol and volatile aroma compounds, allowing them to be collected at gentler temperatures and helping to preserve delicate aromatic molecules that might otherwise be degraded or lost.

The analytical work focused on three complementary layers of the spirits’ chemistry. First, basic physicochemical properties such as alcohol content were measured to assess how efficiently each yeast converted sugars into ethanol. Second, the organic acid composition of the distillates was profiled, since organic acids contribute sourness, complexity, and important ester-forming precursors. Third, the relative abundances of volatile aroma compounds were compared, capturing the esters, higher alcohols, and other molecules that dominate the human perception of flavor in distilled spirits. Together, these measurements allowed the study to map how each microbial strategy translated into a distinct chemical signature in the glass.

One of the most striking results concerned fermentation performance. Among all the inoculated yeasts, Lachancea thermotolerans produced the highest alcohol content in the main distillate before the standard adjustment step that brings commercial soju to its final bottling strength. This finding aligns with the growing reputation of L. thermotolerans in the beverage industry, where it is increasingly valued for its robust fermentative capacity and its ability to modulate acidity. In the context of distilled soju, a yeast that reliably pushes ethanol yields higher could offer distillers both economic and quality advantages, since fermentation efficiency directly affects the quantity and character of the raw spirit collected.

Paradoxically, the control fermentation, which relied entirely on the microbes naturally present in the nuruk mash, produced the highest relative abundances of total esters and of ethyl acetate, one of the most abundant and influential aroma compounds in distilled spirits. Ethyl acetate contributes fruity, slightly solvent-like notes at moderate concentrations, and its prominence in the control sample suggests that the spontaneous microbial community of nuruk remains a powerful engine of ester formation. This result underscores a central tension in modern traditional-spirit production: wild, uncontrolled fermentation can generate exceptional aromatic richness, but it does so at the cost of consistency, which is precisely what commercial starter cultures are meant to provide.

The mixed-yeast preparation combining Saccharomyces cerevisiae with Torulaspora delbrueckii emerged as perhaps the most distinctive profile in the study. This co-inoculation produced the highest total organic acid content of any treatment, along with elevated relative abundances of total higher alcohols and isoamyl alcohol, a compound associated with malty, banana-like notes. At the same time, the mixed culture yielded only a very weak furfural signal. Furfural, which forms from heat-induced degradation of sugars during distillation, contributes almond-like and caramelized nuances but can also impart harshness at higher levels. A very low furfural signal, combined with enriched organic acids and higher alcohols, suggests that the mixed-yeast approach could produce a rounder, fuller-bodied spirit with fewer sharp edges, offering distillers a way to engineer complexity without relying on uncontrolled fermentation.

Metschnikowia pulcherrima, another of the wild yeasts tested, left its own unmistakable mark on the distillate. Spirits fermented with this organism showed relatively high signals for isoamyl acetate, the ester responsible for banana-like aromas, along with ethyl lactate, which contributes creamy and fruity nuances, and 2-phenylethanol, a rose-scented higher alcohol highly prized in both wine and spirits. This trio of compounds points toward a floral, fruity, and soft aromatic profile, consistent with the growing use of M. pulcherrima as a co-fermenter in winemaking, where it is known to enhance aroma intensity and freshness. Its performance in a distilled soju context demonstrates that the aromatic talents of this yeast survive the distillation process and can be transferred to spirits, not just to wine.

Taken together, the results deliver a clear message: yeast selection is not a minor technical detail in soju production but a primary determinant of the spirit’s chemical identity. Each of the tested yeasts produced a measurably different profile of alcohols, acids, esters, and aroma-active compounds, meaning that distillers can, in principle, choose a starter culture the way a perfumer chooses an ingredient, steering the final product toward fruitiness, fullness, floral character, or clean neutrality. The use of commercially available wild yeasts also offers a middle path between the unpredictability of spontaneous nuruk fermentation and the uniformity of a single brewing strain, preserving some of the aromatic richness of tradition while adding a measure of reproducibility that modern markets demand.

The study also situates Korean soju within a global research trend. Non-Saccharomyces yeasts such as Torulaspora delbrueckii, Lachancea thermotolerans, and Metschnikowia pulcherrima have been the subject of intense investigation in wine, beer, and other fermented foods over the past decade, with researchers documenting their roles in modulating acidity, releasing bound aroma compounds, and reducing undesirable byproducts. Applying these organisms to a traditional Asian distilled spirit made from nuruk mash represents a meaningful extension of that work, bridging old-world fermentation practices and new-world microbial technology. As consumer interest in craft spirits and terroir-driven flavors continues to grow, the ability to tune the aroma of distilled soju through deliberate yeast selection could open new categories of premium products, giving one of Korea’s oldest drinks a scientifically engineered future without abandoning the microbial heritage at its core.

Subject of Research: Effects of commercial wild yeasts on the chemical and aroma profiles of Korean distilled soju produced from nuruk mash

Article Title: Effects of commercial wild yeasts on the chemical and aroma profiles of Korean distilled soju produced from nuruk mash

Article References: Son, E.-S. (2026). Effects of commercial wild yeasts on the chemical and aroma profiles of Korean distilled soju produced from nuruk mash. Food Science and Biotechnology. https://doi.org/10.1007/s10068-026-02294-7

Image Credits: AI Generated

DOI: 10.1007/s10068-026-02294-7

Keywords: distilled soju, nuruk mash, non-Saccharomyces yeasts, Lachancea thermotolerans, Metschnikowia pulcherrima, Torulaspora delbrueckii, Saccharomyces cerevisiae, aroma compounds, organic acids, vacuum distillation, Korean traditional alcohol, fermentation

The Balbiani body, often abbreviated as the B-body or Bb, is a cloud of RNA, proteins, and organelles that appears in the cytoplasm of young oocytes across a striking range of animals, from fruit flies to frogs to humans. Although its existence has been known since the nineteenth century, its exact molecular recipe has remained elusive, particularly outside a handful of classic model organisms. In fish, another related structure called nuage—a diffuse, electron-dense material associated with germ cells—complicates the picture. Both structures are membraneless, which means they cannot simply be plucked out of a cell with conventional biochemical tools, and their components intermingle with the surrounding cytoplasm in ways that frustrate isolation.

To overcome this technical barrier, the team led by Ji Liu, Xi Wu, and Jian Luo of Hainan University turned to laser capture microdissection, a technique that uses a precision laser to excise specific regions from tissue sections mounted on a microscope slide. Working with the Leica LMD7000 system at the Ocean University of China, the researchers carved out the Balbiani body region from oocytes at different developmental stages. Because the nuage proved too difficult to isolate reliably on its own, they instead dissected the nucleus together with the surrounding nuage, creating what they call a nuage-associated nucleus-containing region, or NNR. This pragmatic compromise allowed them to profile the nuage neighborhood while acknowledging that nuclear molecules would inevitably contaminate the signal.

The ultrastructural work, carried out with electron microscopy, established the timeline of events. Nuage material was present from the earliest oogonial stages and persisted throughout oocyte development. The Balbiani body, by contrast, made its appearance specifically in primary oocytes, where it gradually expanded before dispersing its contents into the surrounding cytoplasm. This choreography—formation, expansion, and dispersion—mirrors what has been described in other species and suggests a conserved role in preparing the egg with maternal materials that the early embryo will depend on before its own genes switch on.

On the molecular side, the researchers applied Smart-seq2, a highly sensitive method for sequencing the RNA of individually captured cells or compartments, alongside data-independent acquisition proteomics, or DIA, which quantifies thousands of proteins in a sample by systematically fragmenting all peptides within defined mass windows. Running both technologies on the same microdissected regions allowed the team to compare the transcriptomic and proteomic landscapes directly. The result was revealing: across the different Balbiani body stages, RNA levels changed relatively little, but the protein complement shifted much more dramatically. This decoupling suggests that post-transcriptional regulation—proteins being made, modified, or degraded without corresponding changes in messenger RNA—dominates the maturation of the granule, a theme that resonates with the B-body’s known function as a site of stored, translationally repressed maternal transcripts.

The NNR samples, meanwhile, showed molecular variation that tracked with developmental stage, consistent with the changing composition of the nuage that surrounds the nucleus as oocytes age. The compartment-biased expression of established marker molecules matched what is already known about Balbiani body biology, lending confidence to the microdissection strategy. The authors are careful, however, to flag the limits of the NNR data: because the dissected region includes the nucleus, any functional signals detected there must be treated as exploratory rather than definitive. Nuclear transcripts and proteins inevitably dominate such samples, and disentangling genuine nuage components from nuclear background will require more refined approaches in future work.

Perhaps the most consequential finding of the study is the identification and validation of two candidate Balbiani body markers: Ddx6 and Igf2bp3. Ddx6 encodes a DEAD-box RNA helicase, a class of enzymes best known for remodeling RNA structures and regulating translation, and it has well-documented roles in germ granules across the animal kingdom. Igf2bp3, a member of the insulin-like growth factor 2 mRNA-binding protein family, is an oncofetal RNA-binding protein that has been implicated in RNA localization and stability. In the grouper oocytes, both the messenger RNAs and the proteins corresponding to these two genes localized precisely to the Balbiani body region under the microscope. Even more strikingly, their localization shifted over time in lockstep with the granule’s own life cycle—appearing as the B-body formed, tracking its expansion, and dispersing as the structure dissolved into the cytoplasm.

This stage-dependent co-localization makes Ddx6 and Igf2bp3 valuable new tools for the field. Reliable markers are the currency of organelle biology: they allow researchers to track a structure’s dynamics in live tissue, to test whether candidate interactors genuinely join the compartment, and to compare granule composition across species and developmental contexts. For fish, where the Balbiani body has been far less characterized than in frogs or flies, adding two validated markers to the repertoire represents a meaningful expansion of the experimental toolkit. It also opens the door to functional studies—knockdown or knockout experiments could now ask what happens to oocyte quality and embryonic development when these granule components are removed.

The choice of species matters as much as the methods. The leopard coral grouper, Plectropomus leopardus, is one of the most prized reef fish in Asian aquaculture, and its seed production depends on high-quality eggs whose developmental potential is laid down during oogenesis. Understanding the molecular machinery that packages maternal determinants into the oocyte could eventually inform hatchery practices, broodstock management, and breeding programs. The study was supported by funding from the Sanya Yazhou Bay Science and Technology City project, the National Natural Science Foundation of China’s Regional Innovation and Development Joint Fund, and several Hainan provincial programs, reflecting the region’s investment in marine breeding science.

Beyond aquaculture, the work speaks to a broader question in cell biology: how membraneless compartments, held together by weak multivalent interactions rather than lipid bilayers, achieve precise composition and timing. The Balbiani body is a natural model for this problem, assembling at a specific place and time, carrying a defined cargo, and then dissolving on schedule. By pairing spatially resolved microdissection with dual transcriptomic and proteomic profiling, the grouper study demonstrates a framework that other researchers can adapt to any tissue where membraneless granules are too small or too intermingled for conventional purification. The authors themselves emphasize that the nuage remains the next frontier—its molecular composition, and its possible relationship with the Balbiani body, await isolation methods sharp enough to capture it alone. Until then, the dynamic portrait of Ddx6 and Igf2bp3 tracing the granule’s rise and fall offers a vivid glimpse of how an egg builds the seed of an embryo, one molecular granule at a time.

Subject of Research: Molecular characterization of Balbiani body formation during oocyte development in leopard coral grouper

Article Title: Integrative analysis of Smart-seq2 and DIA proteomics reveals dynamic formation of the Balbiani body during oocyte development in leopard coral grouper (Plectropomus leopardus)

Article References: Liu, J., Wang, G., Zhang, X., Wu, X., Zhang, X., Zheng, T., Wen, X., Wu, X., & Luo, J. (2026). Integrative analysis of Smart-seq2 and DIA proteomics reveals dynamic formation of the Balbiani body during oocyte development in leopard coral grouper (Plectropomus leopardus). BMC Genomics. https://doi.org/10.1186/s12864-026-13414-2

Image Credits: AI Generated

DOI: 10.1186/s12864-026-13414-2

Keywords: Balbiani body, oocyte development, nuage, laser capture microdissection, Smart-seq2, DIA proteomics, Ddx6, Igf2bp3, leopard coral grouper, germ granules, oogenesis, membraneless organelles

The core logic of MIDAS is disarmingly simple. Instead of building plasmids, researchers design overlapping PCR fragments that together encode a complete transcription unit, including a promoter, the variant coding sequence, and a polyadenylation signal. Mutations are introduced through primers at defined target regions, and a secondary overlap-extension PCR stitches the fragments into full-length genes. Each PCR product encodes one defined variant and is transfected into one well of a multiwell plate, so the association between sequence and measured function is preserved by well position alone. Because no ligation, transformation, bacterial culture, plasmid purification, or sequence verification is required, generating 384 variants takes roughly four hours of hands-on time and about 2,000 dollars in reagents, compared with an estimated 192 hours and 20,000 dollars for conventional cloning of the same number of constructs.

The platform comes in several flavors that expand its reach beyond single-site substitutions. Polytemplated monofocal MIDAS, or MIDAS-PM, uses multiple plasmid templates to generate arrays of mutants at one location. Polytemplated polyfocal MIDAS, or MIDAS-PP, extends this to combinations of changes at multiple sites by mixing primary PCR products combinatorially. Most notably, the team developed monotemplated protocols, MIDAS-MM and MIDAS-MP, which work from a single plasmid template. The trick lies in adding unique DNA tags to the 5-prime ends of the outermost primers, so that secondary PCR primers recognize only the assembled primary products and never re-amplify the parental plasmid. Sequencing confirmed that this nested-primer design eliminates template contamination, meaning researchers can begin mutagenizing a protein of interest immediately without first subcloning it into multiple backbone plasmids.

To demonstrate the method, the team took on an ambitious target: a bioluminescent indicator for acetylcholine, a neurotransmitter central to arousal, attention, learning, and memory, and one whose signaling declines early in Alzheimer’s, Parkinson’s, and Lewy body dementias. Fluorescent acetylcholine sensors exist, but they require implanted optical elements to deliver excitation light, which is impractical for moving organs outside the brain. A bioluminescent sensor, by contrast, generates its own light and could report cholinergic activity noninvasively anywhere in the body. The researchers built their prototype, ACh-NeuBI0.1, by inserting split NanoLuc luciferase fragments into OpuBC, a bacterial periplasmic binding protein previously engineered to bind acetylcholine over choline. The initial construct responded to acetylcholine with a modest 22 percent luminescence increase, a starting point that would demand extensive optimization.

MIDAS delivered that optimization in rapid iterative rounds. First, MIDAS-MP screened all 36 combinations of glycine linker lengths connecting the binding protein to the luciferase fragments, identifying a configuration that boosted responsiveness by roughly 50 percent. Next, MIDAS-MM tested all 20 amino acids at a single linker position, revealing that phenylalanine improved the response to about 80 percent. A third round tuned the energetics of luciferase fragment assembly by screening SmBiT variants, finding that a three-residue C-terminal truncation raised the response to approximately 2.5-fold. The team then fused the orange fluorescent protein mScarlet-I to the sensor, enabling resonance energy transfer that shifts a portion of the emission above 610 nanometers, wavelengths that penetrate tissue far more effectively than NanoLuc’s native blue light. At each step, plasmid-based validation confirmed that the PCR-transfected results were reliable.

The deepest round of engineering showcased MIDAS’s capacity for true combinatorial saturation mutagenesis. Guided by computational structure predictions from Chai-1, the researchers selected 16 sites around the acetylcholine-binding pocket and screened every possible amino acid at each, uncovering beneficial mutations at three second-shell positions. Because positions 560 and 610 sit close together in space, the team then used MIDAS-MP to test all 400 combinations of side chains at those two residues. Two winners emerged: variant QAE, named ACh-NeuBI1b, which achieved the highest affinity at 188 micromolar, and variant QED, named ACh-NeuBI1c, which suppressed baseline luminescence to deliver maximum contrast. Strikingly, ACh-NeuBI1c responded to 100 micromolar acetylcholine with a 640 percent signal increase, a 29-fold improvement over the original prototype.

The mathematics of why this matters are sobering. Seven of the nine beneficial amino acid changes discovered in the study required two or three nucleotide substitutions. Under error-prone PCR, a functional clone carrying two specific mutations appears roughly once in every 71 million clones, and some combinations once in 140 million, numbers far beyond what multiwell screening can reach. Deterministic PCR assembly sidesteps this barrier by constructing exactly the variants of interest, one per well, whether that means 20 samples for one site or 400 for two. The approach also outperforms degenerate oligonucleotide libraries, which require oversampling that balloons combinatorially: two sites of saturation mutagenesis would demand more than 10,000 screened samples, roughly 25 times what MIDAS requires.

The team applied the same logic to NanoLuc luciferase itself, performing saturation mutagenesis at all 25 residues lining or influencing the substrate-binding pocket, generating 500 variants in a single MIDAS-MM campaign. A D108N mutation improved photon production across three different substrates, while D108S, which requires two adjacent nucleotide changes and would be vanishingly rare in random libraries, selectively enhanced activity on the aqueous substrate FFz. Beyond engineering, the complete activity matrix enabled sequence-fitness analysis, mapping which active-site positions tolerate substitution and which are intolerable, and revealing substrate-specificity determinants whose locations often defied structural prediction. Because NanoLuc produces too few photons for fluorescence-activated cell sorting, the multiwell format proved essential, connecting MIDAS to the gold-standard analytical assays, from luminescence to absorbance to chromatography, that apply to most enzymes.

The biological payoff extended from cell culture into living animals. In primary cortical neurons, ACh-NeuBI1b and ACh-NeuBI1c achieved acetylcholine responses of up to 200-fold while maintaining selectivity over choline and other neurotransmitters. When expressed in mouse liver by hydrodynamic transfection, the optimized indicators detected intraperitoneally administered acetylcholine with signal increases of up to 173-fold, far exceeding the 30-fold response of the pre-optimization sensor, while an unregulated control reporter showed no response. The developers emphasize that improvements measured in mammalian cells, including affinity gains that did not reproduce when proteins were screened as bacterial periplasmic constructs, translated directly into in vivo performance, underscoring the value of screening in the environment where a protein must actually work.

Looking forward, the authors position MIDAS as a data engine for the machine-learning era of protein design. Computational models excel at proposing candidate sequences but cannot reliably identify the single best one, and they remain poorly trained on the dynamic conformational changes underlying enzyme catalysis and biosensor activation. By generating hundreds of precisely defined variants with high-quality functional measurements in mammalian cells within a day, MIDAS can supply exactly the kind of dense sequence-fitness datasets that such models need. The method blurs the traditional boundary between screening and analysis, unifying them in a single workflow that is roughly 48 times faster and ten times cheaper than cloning-based approaches. For laboratories optimizing biosensors, enzymes, or therapeutic proteins in mammalian systems, the bacterial incubator may no longer be the obligatory first stop.

Subject of Research: A cloning-free PCR-based platform for rapid protein engineering and sequence-fitness analysis in mammalian cells

Article Title: Fast analysis and engineering of protein function by microbe-independent deep assembly and screening

Article References: Wu, Y., Wang, P., Liu, L. X., Song, D., Qin, Q., Gao, C., Hageman, M., Kirkland, T. A., Su, Y., & Lin, M. Z. (2026). Fast analysis and engineering of protein function by microbe-independent deep assembly and screening. Molecular Systems Biology, 22(6), 1003-1034. https://doi.org/10.1038/s44320-026-00210-z

Image Credits: AI Generated

DOI: 10.1038/s44320-026-00210-z

Keywords: protein engineering, MIDAS, PCR assembly, mammalian cells, directed evolution, bioluminescent indicator, acetylcholine, NanoLuc, saturation mutagenesis, sequence-fitness analysis, biosensors, Molecular Systems Biology

AMH has long occupied a central place in reproductive endocrinology. Secreted by small growing follicles in the ovary, the hormone serves as a circulating proxy for the size of the resting follicle pool, the reserve that determines both fertility potential and the trajectory toward menopause. In assisted reproduction clinics worldwide, AMH measurements guide ovarian stimulation protocols and counseling about expected response. Yet the hormone’s clinical usefulness in midlife women has remained constrained by two practical problems. First, most existing reference ranges were derived from populations of reproductive-age women, so precise age-specific values for women over 35, and especially those navigating the menopausal transition, have often been lacking. Second, conventional immunoassays can lose analytical sensitivity at the very low concentrations characteristic of late reproductive life, making subtle but clinically meaningful changes difficult to detect and quantify.

To address both limitations, a research team led by Professor Yingying Qin of Shandong University, in collaboration with Professor Ruimin Zheng of the National Center for Women and Children’s Health, drew on a large nationwide cohort of Chinese women aged 35 to 55. By applying an ultrasensitive AMH assay, the investigators were able to measure the hormone with greater precision in precisely the age range where levels approach the floor of detection. This proved particularly important among women aged 44 to 49, a phase the researchers characterize as one of near-depletion, when conventional assays may return values so low that differences between individuals are obscured. The resulting age-specific percentile table, published online on August 17, 2026, in the Chinese Medical Journal, functions as a calibrated chart: a clinician can take a woman’s exact age and her AMH value and locate where she sits relative to the distribution of her peers, seeing at a glance whether her ovarian reserve appears well ahead of or lagging behind the norm for her age.

The percentile framework is more than a descriptive atlas. The team evaluated its clinical utility in a hospital-based care cohort, focusing on women aged 35 to 40 who still planned to conceive. Among these patients, those whose baseline AMH fell below the 10th percentile for their age carried a substantially higher risk of progressing to severely diminished ovarian reserve, defined in the study as an AMH concentration below 0.25 nanograms per milliliter, compared with women whose values sat within the normal percentile range. That distinction matters for fertility counseling. A single AMH reading, interpreted against an age-matched distribution rather than a one-size-fits-all cutoff, can give a woman in her late thirties a quantitative, individualized signal about how quickly her reserve may be declining, informing decisions about the timing of pregnancy attempts or the pursuit of assisted reproduction.

The study’s ambitions extend beyond fertility. Perimenopausal symptoms, particularly vasomotor complaints such as hot flashes and night sweats, are among the most common and most under-recognized problems in routine midlife care, yet they are typically assessed only through subjective questionnaires. The researchers found that among women aged 35 to 55, lower AMH levels were associated with a greater likelihood of more severe perimenopausal symptoms, with the relationship most evident for vasomotor symptoms. This raises the possibility that AMH could serve as an objective biomarker to complement symptom inventories, helping clinicians stratify patients by the likely intensity of their transition and tailor management accordingly, including consideration of hormone replacement therapy when clinically appropriate. The authors are careful to frame this as an association that supports more integrated care rather than a deterministic prediction; symptom experience varies widely among women with similar hormone profiles, and AMH would supplement, not replace, careful clinical assessment.

A third application tested in the study concerns the diagnosis of menopause itself. Clinically, menopausal status is usually inferred from menstrual history together with hormone measurements, but follicle-stimulating hormone and estradiol can fluctuate dramatically during the transition, sometimes producing ambiguous results. Comparing AMH levels between premenopausal and postmenopausal women, the team observed a cliff-like drop, with mean values of 0.187 nanograms per milliliter in premenopausal participants versus 0.043 nanograms per milliliter in those who were postmenopausal. The magnitude of that separation is striking at the group level, but the hormone’s power to classify any individual woman proved moderate, with an overall discriminative performance of around 0.7 on the area under the receiver operating characteristic curve, a standard measure of diagnostic accuracy in which 0.5 indicates chance performance and 1.0 indicates perfect classification.

That moderate AUC carries an important practical message: AMH by itself is not sufficient as a stand-alone diagnostic test for menopause. Menopause remains, fundamentally, a retrospective clinical diagnosis anchored in twelve consecutive months of amenorrhea. However, the study suggests that extremely low AMH values, measured with an ultrasensitive assay, can still provide useful supportive information in clinically challenging situations, for example when a woman’s menstrual pattern is disrupted by contraception or other conditions and her FSH and estradiol results fluctuate across the diagnostic threshold. In such cases, a near-undetectable AMH reading adds a piece of biological evidence that, combined with the rest of the clinical picture, can help clinicians assess where a patient stands in the transition.

The technical achievement underlying these applications lies in the assay itself. At the low concentrations typical of women in their mid-to-late forties, conventional platforms approach their limits of quantification, and small analytical imprecision can translate into large relative errors. An ultrasensitive assay extends the measurable range downward, allowing the near-depletion phase to be characterized rather than lumped into a single undetectable category. That analytical gain is what makes an age-specific percentile table feasible across the full 35-to-55 span: without reliable measurement at the low end, the lower percentiles for older age bands would be statistically unstable and clinically uninterpretable. The nationwide cohort design further strengthens the reference values by capturing the geographic and demographic breadth of the Chinese population, an important consideration given that AMH distributions can vary across ethnic groups and study settings.

Taken together, the new scale addresses practical clinical gaps across both the late reproductive and perimenopausal stages, and the researchers frame its contribution around three integrated goals. Earlier identification means recognizing women at higher risk of low ovarian reserve sooner, providing quantitative support for fertility planning in later reproductive age. Better management means combining AMH profiling with symptom assessment to support more precise handling of perimenopausal complaints and improve quality of life in midlife. Supportive diagnosis means acknowledging that while AMH alone is not an ideal menopause test, very low values measured ultrasensitively can assist clinicians in assessing menopausal status, particularly in complex cases where other reproductive hormones are difficult to interpret. The work was supported by the National Natural Science Foundation of China, the National Key Research and Development Program of China, and several provincial and institutional funders, reflecting the scale of investment in reproductive aging research in China.

For women’s health practice, the study signals a shift toward treating the menopausal transition with the same quantitative rigor long applied to earlier reproductive life. An age-calibrated AMH reference scale gives clinicians a common language for discussing ovarian reserve with patients in their thirties, forties, and fifties, replacing vague reassurances or alarming single cutoffs with a percentile position that evolves with age. It also opens the door to longitudinal use: repeated measurements plotted against the age-specific curve could, in principle, reveal whether an individual woman’s reserve is declining on a typical trajectory or at an accelerated pace, though the current study establishes the cross-sectional framework rather than validating serial monitoring. As ultrasensitive assays become more widely available, the approach developed by Qin, Zheng, and colleagues may prompt parallel reference studies in other populations, a necessary step before the Chinese percentile tables are applied globally. What the study already demonstrates is that the hormonal signature of reproductive aging can be measured, mapped, and translated into clinical decision support across the full arc of midlife, from fertility planning to symptom management to the supportive assessment of menopause itself.

Subject of Research: Age-specific anti-Müllerian hormone reference values for reproductive aging in Chinese women

Article Title: A new scale for reproductive aging in Chinese women based on ultra-sensitive AMH testing

Article References: A new scale for reproductive aging in Chinese women based on ultra-sensitive AMH testing. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: AMH, anti-Müllerian hormone, ovarian reserve, reproductive aging, menopause, perimenopause, ultrasensitive assay, fertility, vasomotor symptoms, hormone replacement therapy, Chinese Medical Journal, women's health

DNA methylation is one of the body’s most important epigenetic regulatory systems. It involves the addition or removal of methyl groups at cytosine-phosphate-guanine (CpG) sites along the DNA strand, frequently in promoter regions, and these chemical tags act like volume knobs for gene expression, dialing genes up or down without altering the underlying genetic code. Methylation patterns shift dramatically during periods of physiological upheaval, and pregnancy is no exception. Previous research has documented methylation changes in maternal blood across gestation, including gains of methylation in genes involved in morphogenesis and losses in genes promoting maternal-infant bonding. Other studies have tied methylation shifts to complications such as gestational diabetes and preeclampsia. But a critical question lingered: do these pregnancy-induced changes simply wash away after delivery, or do they endure as a kind of molecular memory?

The research team, led by Su Chen and Miranda Johs of the University of Nebraska Medical Center alongside collaborators at the University of Memphis, University of Southampton, University of Michigan, Michigan State University, and the David Hide Asthma and Allergy Research Centre, was designed to answer exactly that question, and to fix two persistent gaps in the literature. First, most prior studies of parous women, those who have given birth, lacked nulliparous controls, women who have never given birth, making it impossible to isolate the effect of childbearing itself. Second, only a single small pilot study had ever examined whether methylation changes persist beyond the immediate postpartum period. The new work addressed both problems by comparing methylation profiles of parous and nulliparous women at two timepoints in young adulthood, ages 18 and 26, which correspond to pre-pregnancy and at least six months postpartum for the mothers in the sample.

The discovery cohort came from the Isle of Wight (IOW) birth cohort, a three-generation study established in the United Kingdom between 1989 and 1990 to prospectively follow the natural history of asthma and allergy. Of 750 female participants, 144 had DNA methylation measured at both ages 18 and 26, and 89 of those women had childbearing history validated through medical records. Twenty-eight had given birth at least six months before the age-26 measurement, while 61 had not. The researchers screened 389,355 CpG sites, the overlap between the Illumina HumanMethylation450 and MethylationEPIC platforms, using a training-testing screening method called ttScreening, which repeatedly splits samples into training and test sets to filter out uninformative sites. Linear regression analyses followed, adjusting for six blood cell type proportions, residual methylation at age 18, body mass index change, smoking categories, and socioeconomic status.

The result was striking: 184 CpG sites were significantly differentially methylated between parous and nulliparous women after correction for multiple testing. To guard against a cohort-specific fluke, the team pursued replication in an entirely independent population, the ELEMENT cohort from Mexico City, a multi-generation birth cohort of Hispanic women followed for nearly three decades. In the ELEMENT subsample of 54 women, 105 of the 184 CpGs showed regression coefficients pointing in the same direction as the Isle of Wight findings, and 13 of those were statistically significant. Seven of the replicated sites showed lower methylation in parous women and six showed higher methylation, a pattern consistent across both cohorts despite their differences in ethnicity, geography, and age structure. The 13 replicated CpGs mapped to 16 unique genes.

Among those genes, one stands out with particular force: TM2D3, a gene previously implicated in late-onset Alzheimer’s disease through exome-wide association analysis. In the Isle of Wight data, methylation at cg01537571, a site within a CpG island near the transcription start site of TM2D3, was significantly lower in parous women in both cohorts, and methylation at this site correlated positively with gene expression. When the team examined RNA sequencing data from blood samples at age 26, TM2D3 was the only gene among the 16 that showed significantly lower expression in parous compared with nulliparous women. This convergence is provocative given that increased parity has independently been identified as a risk factor for Alzheimer’s disease, hinting that the methylation change at TM2D3 could be one thread connecting reproductive history to neurodegenerative vulnerability.

Another gene of interest is SEMA3A, which encodes semaphorin 3A, a guidance cue signaling molecule that directs developing neurons during embryogenesis and continues to regulate neuroplasticity in adulthood. SEMA3A showed decreased methylation in parous women and appeared in the study’s gene-disease network linked to neoplasms, mental disorders, and reproductive conditions. The gene plays paradoxical roles in cancer: it promotes tumor progression in hepatocellular carcinoma by enhancing proliferation, migration, and invasion, yet acts as a tumor suppressor in head, neck, and breast cancers, where it boosts antitumoral M1 macrophage proliferation and recruits CD8-positive T cells and natural killer cells to repress tumor growth. The authors speculate that lower methylation near the SEMA3A transcription start site in parous women could conceivably relate to their reduced breast cancer risk, though they caution that whether blood methylation mirrors changes in breast tissue remains unknown.

Beyond these two headline genes, the disease association analysis using the DisGeNET database painted a broader picture. Ten of the 16 genes, including SEMA3A, AKAP13, SLC15A2, DOK2, ADAMTS17, ADARB2, NUP37, PLD5, TPK1, and SLC15A3, were most notably linked with neoplasms, mental disorders, and nervous system diseases such as Parkinson’s disease, schizophrenia, and autism, along with substance use disorders. The breadth surprised even the researchers, who had anticipated connections to reproductive cancers but had not expected substance abuse disorders to surface. Gene expression data confirmed that methylation at cg19086905 (ADAMTS17) and cg13375690 (PDE7A) was negatively associated with expression of those genes, providing functional support for at least some of the methylation signals. Notably, several of the identified genes are expressed in the brain, including TM2D3, SLC15A2, PLD5, ADARB2, and PARPBP, while MAP1LC3C, NUP37, and ADAMTS17 are expressed in breast tissue, tissue-specific patterns that may matter enormously for translating blood-based findings into disease mechanisms.

The study’s design carries genuine strengths. Comparing parous women at pre- and post-pregnancy timepoints against matched nulliparous controls in both discovery and replication cohorts is rare, and the six-month minimum postpartum window avoided the transient methylation fluctuations that may have confounded earlier work sampling just two to four days after delivery. Replication across a 98 percent White UK cohort and an all-Hispanic Mexican cohort supports both internal validity and generalizability. Yet the authors are candid about limitations: the parous groups were small, with 28 women in the Isle of Wight and roughly 20 in the ELEMENT analysis; nulliparous status relied on medical records; the duration required for postpartum methylation to stabilize is unknown; and blood serves only as a proxy for changes in other tissues, with roughly half of the identified genes not routinely expressed in whole blood. Factors that change during pregnancy, such as smoking or income, could not be assessed in nulliparous women at comparable times.

Even with those caveats, the implications are considerable. If the methylation changes observed in blood leukocyte DNA reflect parallel epigenetic remodeling in tissues such as the breast or brain, they could furnish a plausible mechanism for how childbearing reshapes a woman’s disease risk across her lifetime, offering a molecular bridge between reproductive history and outcomes ranging from cancer to Alzheimer’s disease. The researchers emphasize that further studies are needed, particularly cohorts with follow-up into later life, to determine whether these parous-related methylation signatures actually predict health outcomes rather than merely accompanying them. For now, the study adds a compelling piece to a growing body of evidence, from brain imaging traces of childbirth to parity-linked dementia risk, that motherhood writes itself into the body in ways science is only beginning to read, one methyl group at a time.

Subject of Research: Persistent effects of childbearing on blood DNA methylation in women

Article Title: Assessing the effect of childbearing on blood DNA methylation through comparison of parous and nulliparous females

Article References: Chen, S., Johs, M., Karmaus, W., Holloway, J. W., Kheirkhah Rahimabad, P., Goodrich, J. M., Peterson, K. E., Dolinoy, D. C., Arshad, S. H., & Ewart, S. (2024). Assessing the effect of childbearing on blood DNA methylation through comparison of parous and nulliparous females. Epigenetics Communications, 4(1), Article 2. https://doi.org/10.1186/s43682-024-00025-9

Image Credits: AI Generated

DOI: 10.1186/s43682-024-00025-9

Keywords: DNA methylation, pregnancy, childbearing, epigenetics, parous, nulliparous, CpG sites, TM2D3, SEMA3A, Alzheimer's disease, breast cancer, Isle of Wight cohort

The proposal, published in the journal Discover Biotechnology, builds on a growing body of work in plant-derived edible vaccines. The idea is deceptively simple: genetically engineer a plant to produce an antigen, a harmless fragment of a pathogen that the immune system can learn to recognize, and then have people or animals eat the plant tissue. When the antigen reaches the gut, it encounters the gut-associated lymphoid tissue, or GALT, a dense network of immune cells that guards one of the body’s main entry points for pathogens. Studies in both mice and humans have shown that plant-derived antigens delivered orally can stimulate antigen-specific mucosal IgA antibodies as well as serum IgG, the systemic antibody class measured after conventional immunization. That dual response matters, because mucosal immunity is the first line of defense against the respiratory and intestinal infections that cause enormous global disease burden.

Plant-based delivery also solves some stubborn engineering problems. The plant cell wall acts as a natural capsule, shielding antigenic proteins from the acidic environment of the stomach so they survive long enough to reach immune tissue. Even more striking, freeze-dried plant material expressing vaccine antigens has been shown to maintain antigen integrity at ambient temperatures, eliminating the need for refrigeration during storage and transport. For regions where electricity is unreliable and cold-chain logistics are prohibitively expensive, a vaccine that can sit on a shelf as dried powder is a fundamentally different proposition from one that must be kept chilled at every step.

Among the many plants tested as vaccine factories, duckweeds—tiny aquatic flowering plants of the family Lemnaceae—stand out. They pack protein at up to 40 percent of their dry weight, grow at remarkable speed on inexpensive byproduct substrates, and, crucially for vaccine consistency, are so small and anatomically uniform that antigen levels should be even between individual plants. That uniformity addresses one of the biggest weaknesses of edible vaccines in general: dosage. When a vaccine is eaten, the amount of antigen consumed can vary depending on which part of the plant is eaten and how much, leading to inconsistent immune responses. Researchers are countering this with genetic engineering strategies that drive uniform antigen expression in specific tissues, and duckweed’s near-homogeneous body plan gives it a natural head start.

The evidence in animal models is encouraging. Duckweed engineered to express Interleukin-17B has been used as a mucosal adjuvant against infectious bronchitis virus in chickens. When the M2e peptide of the H5N1 avian influenza virus—a sequence so conserved it appears across essentially all influenza strains—was expressed in duckweed and fed to mice, the animals produced anti-M2e antibodies detectable in blood serum. And in an aquaculture demonstration, zebrafish fed transgenic duckweed carrying the LamB antigen from Vibrio alginolyticus were protected against fish vibriosis: when challenged with five times the lethal dose of the pathogen, 63 percent of the vaccinated fish survived, compared with complete mortality in the unvaccinated control group.

So why has no one scaled this up? The answer lies in duckweed’s biology. Because the plants are aquatic, they need vast expanses of water to grow, and in water-scarce regions that requirement can outweigh the benefits of the vaccine itself. Artificial ponds would have to be built, consuming capital and competing with agriculture and urban development for land. Open-air ponds are also vulnerable to contamination by microalgae and fungi, which can render entire batches unusable and impose heavy financial losses. Finally, harvesting delicate, millimeter-scale plants is logistically awkward, and ponds are often located far from the end users, forcing complex delivery chains that erode the cost advantages that made duckweed attractive in the first place.

The research team’s answer is a vertical farming and recirculation system. Instead of spreading duckweed across horizontal ponds, the design stacks multiple layers of cultivation trays, with the liquid medium held at a shallow depth of roughly five centimeters—enough for duckweed, which floats at the surface, to thrive. Water is pumped to the top tray and flows gravitationally through the lower trays back to a reserve basin at the bottom, continuously recirculating and dramatically cutting water use. According to the authors, this vertical arrangement could raise production yield per unit area by a factor of ten to twenty compared with conventional cultivation. Biomass is collected with a filtration-based harvester that returns the supernatant to the basin, minimizing waste.

Vertical farms have a well-known Achilles’ heel: trays lower in the stack are shaded by those above. The proposed solution is elegantly mechanical. The trays would be made of transparent material and mounted on retractable mechanisms along both horizontal axes, with a programmable algorithm adjusting the retraction amplitude and frequency in response to real-time light sensors, so each layer of duckweed gets its optimal share of sunlight. Solar panels integrated into the structure would help offset the energy costs of pumping and automation. The team also proposes siting these systems on livestock and poultry farms, where animal wastewater could serve as the growth medium—turning a disposal problem into a nutrient source while the duckweed simultaneously purifies the water. Settling ponds can reduce turbidity and microbial load in the wastewater before use, and prior studies indicate that duckweed grown on wastewater does not suffer unsafe microbial contamination. Because the system is enclosed, it also limits environmental contamination and gene flow, aligning with global biosafety regulations; regular monitoring for heavy metals in the waste streams would be needed, with quality control applied across the value chain to meet Good Manufacturing Practice standards.

The envisioned value chain runs from molecular bioengineering to the dinner bowl. A gene encoding a pathogenic antigen is introduced into duckweed, the transformed plants are cultivated at scale in the retractable vertical trays, and the harvested biomass is dried and ground into a powder that can be mixed into food or feed—immunizing and nourishing the consumer at the same time. The capital costs of such a system, especially when fitted with solar panels, sensors, and automated retraction, exceed those of traditional greenhouse production, but they are far lower than bioreactor-based manufacturing, and operating expenses can be trimmed by optimizing light regimes. The authors argue that high yields and low downstream processing costs could give the approach a decisive edge in overall unit economics, with lighting costs and protein yield as the two variables that will ultimately determine competitiveness.

Significant hurdles remain before duckweed vaccines reach the main course. Purification steps are still required before the material can be used as a vaccine, and the regulatory pathway for an edible, plant-grown immunization is uncharted territory in most jurisdictions. Yet the convergence of advances in plant biotechnology—improved promoter optimization, enhanced expression systems, and synthetic biology tools—with a cultivation system designed for scale suggests that the gap between laboratory proof and global health impact may finally be closable. If powdered duckweed vaccines can be grown beside the farms and villages that need them, stored without refrigeration, and eaten rather than injected, one of medicine’s oldest tools could take on one of its most persistent inequities.

Subject of Research: Vertical farming of transgenic duckweed for scalable production of edible vaccines

Article Title: Vertical farming of duckweed: advancing edible vaccine production for global health

Article References: Grieš, M., Rosputinský, M., Mećava, M., & del Valle, A. C. (2025). Vertical farming of duckweed: advancing edible vaccine production for global health. Discover Biotechnology, 2(1), Article 14. https://doi.org/10.1007/s44340-025-00018-x

Image Credits: AI Generated

DOI: 10.1007/s44340-025-00018-x

Keywords: duckweed, edible vaccines, vertical farming, mucosal immunity, transgenic plants, Lemnaceae, cold chain, wastewater, antigen expression, global health, plant biotechnology, aquaculture