Biology subjects

Gupta, P.

Publications and source records attributed to Gupta, P..

At least 19 recordsLinked to original sources

bioRxiv2026

Iron as a critical determinant for survival, oxidative stress, and infectivity of Leishmania parasites

Iron is an essential for all living organisms, including humans, and is present in RBCs & other iron-binding proteins that play crucial roles in oxygen transport, DNA synthesis & electron transport. Similarly, iron also plays a critical role in parasite biology. The unicellular Leishmania parasites scavenge essential iron inside host macrophages. We investigated the impact of iron on parasite growth, apoptosis, oxidative stress, and the expression of proteins involved in ROS, thiol metabolism, Fe-S clusters, and infectivity in macrophages. Our findings revealed that optimum iron concentrations promote parasite proliferation, whereas higher levels induce toxicity, leading to DNA fragmentation and caspase-like activity. Iron supplementation enhances parasite tolerance to oxidative stress by decreasing intracellular ROS and increasing total intracellular thiol content in a concentration-dependent manner. However, iron depletion increases ROS levels and decreases total thiol content in a concentration-dependent manner, which disrupts redox homeostasis. To support our finding, we further exposed these treated parasites to menadione, which supports that iron-supplemented cells can better survive under oxidative stress than the depleted cells. Furthermore, extra iron availability increases the expression of key metabolic proteins, Fe-S clusters (Nbp35, Frataxin, IScU) and thiol metabolism (TryS, TryR, CS, SAT, cTXN, cTXNPx), while under iron-depleted conditions these proteins were downregulated except frataxin. In vitro infectivity demonstrated that iron supplementation enhances parasite infectivity, while iron depletion impairs percentage infectivity. Our study demonstrates the roles of iron in the survival, redox regulation, and virulence of L. donovani, which can correlate with the role of cellular iron homeostasis in VL pathogenesis.

microbiology↗

bioRxiv2026

hnRNPA1 Remodels SARS-CoV-2 Genomic G-Quadruplexes through Its UP1 Domain

The single-stranded RNA genome of SARS-CoV-2 contains several predicted G-quadruplex (G4) forming sequences. Pharmacological stabilisation of these structures has been shown to suppress viral replication, suggesting that their resolution may be required for efficient viral replication. However, the roles of host proteins in stabilization or destabilization of G4 structures have remained largely unexplored. Human hnRNPA1, one of the most abundant RNA-binding proteins in eukaryotic cells, is a G-quadruplex unfolding protein and has independently been identified as a SARS-CoV-2 genome-binding protein. However, whether it acts on the G-quadruplexes within the SARS-CoV-2 genome has not been examined. Here, we screened eight putative G-quadruplex forming sequences conserved across SARS-related coronaviruses and identified two sequences that fold into G4 structures, confirmed by circular dichroism (CD) spectroscopy and thioflavin T (ThT) fluorescence-based experiments. Furthermore, we showed that the UP1 and UP1+RGG domains of hnRNPA1 specifically bind and unfold these two viral RNA G-quadruplexes. Both UP1 and UP1+RGG unfold viral RNA G-quadruplexes to similar extent, indicating that recognition and unfolding of these viral G-quadruplexes are mediated mainly by the UP1 domain of hnRNPA1. Together, these findings identify hnRNPA1 as a host factor capable of binding and remodelling SARS-CoV-2 genomic RNA G-quadruplexes and suggest a mechanism by which the virus may exploit an abundant host protein to resolve G4 structures within its genome.

biophysics↗

bioRxiv2026

C-terminal domain of SARS CoV 2 nucleocapsid protein binds transcriptional regulatory sequence RNA with high affinity and forms higher-order assemblies

The SARS-CoV-2 nucleocapsid (N) protein is a 419-amino-acid modular protein essential for viral RNA replication, genome packaging, and host-cell modulation. It contains two structured domains, the N-terminal RNA-binding domain (NTD) and C-terminal dimerization domain (CTD), separated and flanked by three intrinsically disordered regions. Although the CTD mediates N-protein dimerization and contributes to RNA binding, the molecular basis of its interaction with RNA remains poorly understood. Here, we characterized CTD binding to transcriptional regulatory sequence (TRS) RNAs from the SARS-CoV-2 genome using NMR spectroscopy, isothermal titration calorimetry, size-exclusion chromatography, and in vitro phase-separation assays. The CTD bound structured stem-loop TRS RNA with higher affinity than the canonical RNA-binding NTD. NMR chemical-shift perturbation experiments indicated formation of higher-order molecular assemblies upon titration of the CTD with longer, structured stem-loop TRS RNA, consistent with size-exclusion chromatography results. The CTD also underwent RNA-induced phase separation in vitro, with stem-loop TRS RNA promoting markedly more robust condensation than single-stranded RNA sequences. Together, these findings establish the CTD as a high-affinity, structure-selective RNA-binding domain with a preference for stem-loop TRS RNA. The results further show that CTD-TRS interactions promote higher-order assembly and phase separation, providing mechanistic insight into N-protein-mediated viral genome packaging and nucleocapsid assembly.

biochemistry↗

bioRxiv2026

An ancestral mitochondrial DNA insertion disrupts RNAi and enables persistence of a novel mycovirus in Cryptococcus neoformans

RNA interference (RNAi) is a widely conserved genome-defense mechanism that protects eukaryotes against viruses and transposable elements. We previously showed that RNAi loss can lead to hypermutation and antifungal drug resistance in the human fungal pathogen Cryptococcus neoformans, illuminating the potential clinical relevance of this pathway. In this study, we identified another function of RNAi in C. neoformans: mycovirus restriction. By screening known RNAi-deficient C. neoformans isolates, we identified a novel dsRNA mycovirus of the Orthototiviridae family, which we named CnTV1. We subsequently detected CnTV1 in three additional RNAi-deficient isolates. All CnTV1-positive isolates shared an ancestral nuclear mitochondrial DNA segment (NUMT) insertion that disrupts the gene encoding Argonaute (Ago1). Restoration of RNAi in sexually produced zygotes efficiently eliminated CnTV1. RNAi rescue by CRISPR-Cas9-mediated allele exchange eliminated CnTV1 during vegetative growth, further demonstrating RNAi is sufficient for mycoviral control. Loss of the RNA helicase Ski2 or the exoribonuclease Xrn1 increased CnTV1 abundance, revealing RNAi-independent restriction of the virus. Restoration of RNAi cleared the virus even in the absence of Ski2 or Xrn1, indicating a dominant role for RNAi in antiviral defense. To investigate the biological implications of CnTV1 infection, we developed a cytoplasmic-mixing approach and generated isogenic strain pairs in an RNAi-deficient background that differ only in viral infection status. Leveraging these strains, we show that CnTV1 infection leads to coordinated, low-magnitude transcriptomic changes and that strains lacking the mycovirus were moderately less virulent in a murine infection model. Our findings reveal that RNAi serves as a dominant antiviral defense system in C. neoformans and suggest that naturally occurring RNAi deficiency may be more prevalent than previously appreciated. This work highlights mycovirus persistence as an important consequence of RNAi loss in this WHO-designated critical priority fungal pathogen.

genetics↗

bioRxiv2026

Cryptic Genes as Building Blocks of Antimicrobial Resistance: Evidence from a Novel Marine Imipenem-Hydrolyzing Metallo-β-Lactamase

The emergence of antimicrobial resistant pathogens has led to increase in mortality rate, longer hospital stays, and need for complex and expensive treatment regimens. However, it is not very clear from where does the bacteria get the building blocks to develop resistance to these synthetic antimicrobial agents which have been integral part of modern medicine. One hypothesis is that non-coding cryptic genes without any vital function or positive contribution to fitness act as versatile endogenous genetic repertoires that help in bacterial adaptation and evolution of new phenotypes when conditions change. This study reports the presence of putative cryptic metallo beta lactamase gene in the marine bacterium Cytobacillus oceanisediminis isolated from the Gulf of Mannar, India. Using in silico, in vitro and in vivo assay, we demonstrate that putative cryptic metallo beta lactamase encodes an enzyme that can cleave only imipenem, has two characteristic domains: HXHXDH and TPGH, but does not share homology with the other reported Metallo-beta-Lactamase genes. The findings underscore the importance for comprehensive screening of putative cryptic antimicrobial resistance genes to fully address and mitigate the challenges posed by antimicrobial resistance.

microbiology↗

bioRxiv2026

Tumor-derived SAA1-TLR4 signaling drives tumor-to-muscle communication in pancreatic cancer cachexia

Cancer cachexia limits treatment tolerance and survival in pancreatic ductal adenocarcinoma (PDAC), yet the tumor-derived signals driving tissue dysfunction remain poorly understood. Here, we identify serum amyloid A1 (SAA1) as a mediator of tumor-to-host communication acting through Toll-like receptor 4 (TLR4). Tumor-derived SAA1 was elevated in human PDAC and in a mouse PDAC model and disrupted both myofiber and muscle stem cell (MuSC) homeostasis. Genetic reduction of tumor-derived SAA1 uncoupled tumor progression from host wasting, preserving muscle mass and function and prolonging survival without affecting primary tumor growth. Mechanistically, SAA1-TLR4 signaling drove multicellular remodeling of the skeletal muscle microenvironment. Therapeutic TLR4 inhibition after cachexia onset restored muscle mass, function and MuSC abundance and prolonged survival independently of tumor growth. Conservation of SAA1-TLR4 signaling in human skeletal muscle identifies a therapeutically actionable tumor-host pathway and demonstrates that host deterioration can be targeted independently of tumor progression.

cancer biology↗

bioRxiv2026

Cysteine reduction rather than methionine restriction promotes immunometabolic fitness to confer survival benefit in aging

Elevated sulfur-containing amino acids (SAA; methionine and cystine) are linked to human mortality. While methionine restriction (MR) extends lifespan in animals, it is often ignored that traditional longevity-promoting MR diets also lack cystine. When dietary cystine is eliminated, host redirects substrates from methionine cycle and generates cysteine via cystathionine {gamma}-lyase (CTH) in transsulfuration pathway, leaving it unclear which SAA controls aging. Here, we show that selective cysteine restriction in Caenorhabditis elegans enhanced lifespan and stress survival independently of methionine. Notably, cysteine-free diets (methionine-replete or restricted) in Cth-deficient mice induced pro-metabolic effects, whereas MR with normal cysteine was ineffective. Sustained 70% cysteine reduction in aged Cth-/- mice reprogrammed the immunometabolic axis, conferring healthspan benefits. Thus, lowering cysteine while keeping the methionine pool intact enhances healthy lifespan.

physiology↗

bioRxiv2026

Solving High-Dimensional Population Balance Equations via Dynamics-Preserving Autoencoders

High-dimensional population balance equations (PBEs) provide a natural framework for modeling heterogeneous cell populations, but their direct numerical solution becomes computationally prohibitive when the internal state space contains many molecular variables. We propose a hybrid mechanistic-machine learning framework for reducing and simulating PBEs defined over high-dimensional intracellular coordinates. The cell population is described by a number density n(x, t), where x [isin] [R]N represents gene and protein states associated with macrophage activation. A dynamics-preserving autoencoder maps this state space to a low-dimensional latent coordinate z [isin] [R]d, with d << N, while retaining key qualitative features of the underlying gene regulatory network, including attractor structure and multistability. Mechanistic information from the original regulatory dynamics is used to construct interpretable drift and diffusion terms for the reduced latent-space PBE. The reduced PBE is solved using a stochastic Lagrangian particle representation, in which particles evolve according to stochastic differential equations (SDEs) corresponding to the latent drift and diffusion fields. The resulting latent-space solution is subsequently decoded and propagated back into the original state space to recover physically interpretable cellular dynamics. We demonstrate the framework on macrophage polarization under cytokine-dependent regulation, including gene knockout perturbations. Overall, the proposed framework provides a computationally tractable and mechanistically interpretable route for integrating single-cell genomic data with population balance models of cell-state dynamics.

systems biology↗

bioRxiv2026

The Critical Role of Pdyn-Lineage Enteric Neurons in Colonic Motility and Visceral Interoception

Summary/Abstract Sensory neurons play well defined roles in the regulation of intestinal motility, digestion, and interoception, but transcriptional dissection of intrinsic and extrinsic sensory neurons innervating the intestines has been challenging because these cells share many genetic markers. However, we cross-referenced transcriptional profiles of intrinsic and extrinsic intestinal neurons and the cells that surround them and identified Pdyn, the gene encoding Prodynorphin, as a marker of putative sensory enteric neurons of the mouse intestines. In a Pdyn lineage-reporter mouse, we identified labeled cells in the myenteric and submucosal plexuses of the large intestine, in contrast to their sparse presence in the dorsal root or nodose ganglia. In dissociated cell culture, these neurons mostly display a phasic firing pattern, discharging one or two action potentials at the onset of a depolarizing current pulse, followed by a prompt cessation of firing despite continued current injection, as assessed by whole-cell patch-clamp recordings. Optogenetic activation of Pdyn-lineage neurons propels stool in ex vivo colons, and in untethered and mobile mice, optogenetic stimulation of the proximal colon induces freezing behaviors and orbital tightening, suggesting interoceptive behaviors, without significant stool output differences. Together, these findings suggest that activation of Pdyn-lineage enteric neurons regulates motility and visceral interoception.

neuroscience↗

bioRxiv2026

Saturation-seq integrates single-cell saturation genome editing and RNA-seq to quantify NFE2L2 (NRF2) variant effects

Interpreting the functional consequences of variants remains one of the central unsolved problems in genomics and clinical genetics. Compounding this, most existing approaches rely on reductive, one-dimensional proxies such as cell growth to score variant effects, which can be a poor substitute for the rich, multidimensional phenotyping that is ultimately needed to understand how variants alter biology. This is especially true for variants known to act through gain-of-function/neomorphic mechanisms. We developed Saturation-seq, a high-throughput platform that combines saturation genome editing with single-cell DNA and RNA profiling to systematically map variant effects. Using CRISPR-based editing in a barcoded haploid cell line, we install hundreds of variants directly into endogenous genomic loci, testing them in multiplex and preserving the native coding and regulatory context. Single-cell amplicon and transcriptome sequencing enables direct linkage of each genomic edit to its transcriptional impact. We apply Saturation-seq to comprehensively characterize 230 variants in the recurrently mutated N-terminal region of NFE2L2 (NRF2), a master regulator of oxidative stress and an oncogene mutated in lung cancer. We define variant function with disruption scores computed from misregulation of known NRF2 targets in single-cell transcriptomes; scores separate pathogenic/benign truthset variants with >90% accuracy and enabled interpretation of TCGA and TRACERx patient tumor data, as well as a rare NFE2L2 germline variant linked to a developmental syndrome. Thus, we establish a broadly applicable high-resolution single-cell variant-to-function platform with a rich phenotypic readout.

genomics↗

bioRxiv2026

KBase Research Agent: Automated Multi-Agent Workflow Construction for Reproducible Genome Analysis

Constructing multi-step bioinformatics workflows, from read quality control through genome assembly to functional annotation, requires expertise in both biology and computational tool selection, creating a bottleneck for scalable and reproducible analysis. We present the KBase Research Agent, a multi-agent system for automating such workflows within the DOE Systems Biology Knowledgebase (KBase). Given a set of sequencing reads and a research objective, the agent constructs an analysis plan grounded in KBase documentation and a Knowledge Graph (KG) of the KBase application catalog, then selects, parameterizes, validates and executes appropriate KBase applications to carry out the workflow. The resulting analysis is preserved as a reproducible KBase Narrative. We evaluate the systems planning and execution quality against ground truth constructed from reference workflows derived from peer-reviewed Microbiology Resource Announcements. We further apply the agent to 100 previously unanalyzed bacterial isolate genomes from the JGI IMG/M database, where it autonomously performed read quality control, genome assembly, taxonomic classification with GTDB-Tk, and downstream analysis producing annotated genomes, reproducible Narratives, and draft manuscripts without human intervention. Across these experiments, the KBase Research Agent demonstrates the feasibility of domain-grounded, end-to-end scientific workflow automation in a production bioinformatics platform.

bioinformatics↗

bioRxiv2026

Molecular mechanism for the functional divergence between cohesin paralogs during meiosis

The key features of meiosis that enhance evolutionary success compared to mitosis are the processes of recombination and independent assortment during chromosome segregation, which help cells adapt to changing environments. Cohesins play critical roles in both these processes and have evolved specialized paralogs that are essential for meiotic chromosome dynamics. Studies so far have elucidated the roles of these meiotic cohesins but it is still unclear why the original mitotic proteins could not serve these evolved functions. In this study, we identify the mechanistic steps that are lost during meiosis when the mitotic counterparts replace the meiotic cohesins in Schizosaccharomyces pombe. The meiotic cohesin subunit Rec8REC8 has evolved multiple unique features that differentiate it from its mitotic paralog Rad21RAD21. Although ectopic expression of Rad21 in meiosis allows its chromatin enrichment, it fails to support reductional separation of chromosomes, initiation of recombination and protection of cohesion in anaphase I, resulting in catastrophic segregation errors. In contrast, the meiotic cohesin regulatory subunit Rec11STAG3, has only one major function of initiation of recombination, which is expectedly hampered in its absence. We show that although the mitotic paralog Psc3STAG1/2 is highly enriched at the cohesin-rich chromosomal axes, it is unable to recruit the downstream activators required for the induction of double-strand breaks. Our work systematically demonstrates the minimal functions that were necessary for the molecular evolution of these paralogs and explains the mechanisms that led to these adaptations.

molecular biology↗

bioRxiv2026

In silico investigation of alternative splicing of microexons in human peripheral tissues

Microexons are highly conserved fragments of exons ranging from 3 to 51 nucleotides (nts), representing a precise but poorly understood layer of post-transcriptional regulation outside the central nervous system. While their role in neuronal development is well documented, their behavior in peripheral tissues remains largely uncharacterized. In this study, we utilized VAST-TOOLS to perform a comprehensive meta-analysis of alternative microexon splicing across independent transcriptomic datasets spanning hepatic, pulmonary, renal, and colonic tissues. By comparing the diseased and wild-type (WT) profiles, we identified a robust set of differentially spliced microexons (DSMs) unique to each disease. Our findings suggest that microexon dysregulation in the liver, lung, kidney and colon may not be a primary driver of specific diseases, but rather a signature of a broader collapse in cellular splicing homeostasis. We propose that this phenomenon of differential splicing, particularly within critical hub proteins, fundamentally compromises protein interaction networks, thereby priming the cell for the diverse phenotypic failures observed across chronic disease states.

genomics↗

bioRxiv2026

Sleep Spindle-Locked Targeted Memory Reactivation Enhances Declarative Memory Consolidation

Study ObjectivesSleep spindles are implicated in memory consolidation. Yet direct evidence linking spindle dynamics to declarative memory outcomes remains limited. We thus tested whether targeted memory reactivation (TMR) time-locked to sleep spindles enhances declarative memory, and whether the temporal organization of stimulated spindles-trains versus isolated events-is selectively associated with distinct memory outcomes. MethodsTwenty-eight healthy young adults learned image locations from two categories (animals, clothing) in a grid, each paired with a distinct auditory cue. During overnight NREM sleep, one cue was replayed time-locked to spindles detected in real-time using a closed-loop system (TMR condition); the other served as the non-reactivated control (No-TMR condition). Category-cue assignment was counterbalanced. Post-sleep recall, recognition accuracy, and movement time were assessed. ResultsRecall accuracy was significantly higher in the TMR than the No-TMR condition (93.96% vs. 90.61%, p = .024), whereas recognition accuracy (p = .139) and movement time (p = .651) did not differ. Stimulation intensity within spindle trains correlated with the TMR effect on recall (Spearman {rho} = .531, p = .004), whereas the proportion of isolated spindle stimulations correlated with the TMR effect on recognition ({rho} = .563, p = .002). Cross-associations were not significant. ConclusionsSpindle-locked TMR enhances recall-based declarative memory retention. The selective association between spindle temporal clustering and memory outcomes suggests that train-embedded and isolated spindles support different aspects of memory consolidation, highlighting spindle temporal context as a functionally relevant dimension of sleep-dependent memory processing.

neuroscience↗

bioRxiv2026

Endocrinology shadows ecology: Characterization of Markhor (Capra falconeri heptneri) reproductive cycles through non-invasive hormone assessment

Markhor (Capra falconeri) is a charismatic, threatened, large, high-altitude bovid found in parts of central and south Asia. The species faces threats such as habitat loss, hunting, poaching, livestock competition, hybridisation, and disease, yet research on wild populations is challenging. Various biological aspects, including surveys, diet, population dynamics, interactions with livestock, hybridisation, and disease, have been studied locally, along with behavior and reproductive biology, but details such as pregnancy, oestrus, and parturition timing remain unestablished. We conducted the first systematic, detailed, and fine-scale characterization of the reproductive steroid profiles of two males and five female markhors (Capra falconeri heptneri) in a captive population at the Padmaja Naidu Himalayan Zoological Park (PNHZP), West Bengal, India. We collected weekly fecal samples, standardized and validated measurements of progesterone (fP4M) and testosterone (fTM) metabolites, and conducted reproductive profiling to assess reproductive stages in both sexes. Analyses of annual fP4M and fTM data from male and female markhor individuals showed similar profiles and synchronicity, with individual variation, and peaks and baselines were evident for both hormones. In both sexes, significantly higher hormone titres were observed during the sexually active and inactive phases. Non-invasive measurement of reproductive hormones accurately reflected ovarian function in females, helping establish mating, gestation, and parturition timelines in female markhors and determine the breeding season in males. These approaches support husbandry and breeding management by identifying optimal pairing, diagnosing pregnancy, and predicting parturition in both captive and wild populations. When applied correctly, these tools could greatly aid population monitoring of other endangered species across high-altitude regions worldwide.

physiology↗

bioRxiv2026

Environmental chemical mixtures reprogram mammary epithelial development to epigenetic states associated with breast cancer

Environmental exposures occur as complex mixtures, yet the mechanisms by which they alter human tissue development and confer cancer susceptibility remain poorly defined. Here, we establish a physiological 3D human breast organoid high-content screening (3D-HCS) platform that quantitatively links exposure-induced developmental disruption to cancer-relevant cell states. Screening of structurally diverse environmental chemicals as single agents but also as mixtures reveals heterogeneous but reproducible developmental responses, indicating that distinct environmental chemicals perturb different aspects of mammary morphogenesis. Focusing on bisphenols, we show that a physiological-dose mixture (BPA, BPS, and BPF; 3BPX) disrupts organoid development and induces transcriptional programs characterized by extracellular matrix remodeling, epithelial plasticity, and partial epithelial-mesenchymal transition (EMT), accompanied by widespread DNA methylation remodeling. The 3BPX exposure signature maps to ER luminal breast cancers and is preferentially enriched in invasive lobular carcinoma (ILC), a subtype characterized by epithelial plasticity and stromal interaction. Genome-wide DNA methylation profiling revealed that the 3BPX-associated methylation signature is detectable in primary tumors and adjacent normal tissue and is enriched in Normal-like tumors and ILC. Together, these findings demonstrate that environmentally relevant chemical exposures induce heterogeneous developmental perturbations that converge on conserved transcriptional and epigenetic programs associated with epithelial plasticity, tissue remodeling, and cancer susceptibility. These exposure-induced states persist as molecular "scars" that map to defined epithelial and stromal compartments in human breast cancers, supporting a model in which developmental exposures establish a field of increased oncogenic potential. Summary StatementUsing a human breast organoid high-content screening platform, we show that real-world chemical mixtures induce heterogeneous developmental disruption that converges on conserved EMT, wound-response, and epigenetic programs. These exposure-induced states persist as molecular "scars" and map to specific epithelial and stromal compartments in human breast cancers, particularly invasive lobular carcinoma, suggesting a developmental origin of cancer susceptibility.

cancer biology↗

bioRxiv2026

Meiotic cohesin paralogs govern cell survival by exhibiting flexibility in partner choice

Erroneous loading of the ring-shaped cohesin complex, especially at centromeres, cause chromosomal segregation defects in both mitosis and meiosis. Mitotic cohesin subunits of this complex, either get replaced or co-exist with their meiotic paralogs during meiosis and also in certain cancers. However, it is unclear whether meiotic paralogs can partner mitotic subunits to form hybrid complexes in somatic cells and if there are any functional consequences on cancer progression. Here, we provide a conceptual framework for the principles of cohesin complex assembly involving non-canonical subunits in proliferating Schizosaccharomyces pombe. We show that chromosome loading, segregation fidelity and cellular proliferation are critically affected by the composition of the available cohesin complexes. We find stark differences in the ability of the meiotic kleisin subunit Rec8 to support robust centromere loading, irrespective of its partner, when compared to the canonical mitotic paralog Rad21. Such variations in cellular growth can be explained by different dwell times of these cohesin complexes on the chromosomes as determined by single-molecule tracking and altered chromatin enrichment. We also discover a unique feature of Rec8, in stabilizing chromatin-bound hypomorphic cohesin mutants that aid in cell survival under restrictive conditions. Overall, we highlight the flexibility of meiotic cohesins in restoring function, albeit at a fitness cost, in the presence of inactivating cohesin mutations. Such imbalances could be exploited by cancers to aid cell survival, but at the expense of increased aneuploidy and genomic instability.

molecular biology↗

bioRxiv2026

BioReason-Pro: Advancing Protein Function Prediction with Multimodal Biological Reasoning

Protein function annotation is fundamental to understanding biological mechanisms, designing therapeutics, and advancing biomedical research. Current computational methods either rely on shallow sequence similarity or treat function prediction as isolated classification tasks, failing to capture the integrative reasoning across sequence, structure, domains, and interactions that expert biologists perform to infer function. We introduce BioReason-Pro, the first multimodal reasoning large language model (LLM) for protein function prediction that integrates protein embeddings with biological context to generate structured reasoning traces. A key input into BioReason-Pro is the set of GO term predictions made by GO-GPT, our autoregressive transformer that captures hierarchical and cross-aspect dependencies of GO terms. BioReason-Pro is trained via supervised fine-tuning on synthetic reasoning traces generated by GPT-5 for over 130K proteins and further optimized through reinforcement learning. It achieves 73.6% Fmax on GO term prediction and an LLM judge score of 8/10 on functional summaries, substantially outperforming previous methods. Evaluations with human protein experts show that BioReason-Pro annotations are preferred over ground truth UniProt annotations in 79% of cases. Remarkably, BioReason-Pro predicted a novel interaction partner for the renal cancer biomarker RCDG1, which we confirmed in the lab by co-immunoprecipitation. In other binding-partner predictions, its per-residue attention localized to the exact contact residues resolved in cryo-EM structures. Together, GO-GPT and BioReason-Pro establish a framework for protein function prediction that combines precise ontology modeling with interpretable biological reasoning.

molecular biology↗