Biology subjects

Corso, G.

Publications and source records attributed to Corso, G..

17 recordsLinked to original sources

bioRxiv2026

BoltzMol-1: Towards Reliable Virtual Screening for Fast and Cost-Effective Hit Discovery

We present BoltzMol-1, a small-molecule hit discovery pipeline, centered on an optimized version of Boltz-2, explicitly adapted for prospective discovery. Reliable hit discovery that generalizes across target classes (rather than only the well-characterized families that dominate existing ligand data) would broaden the range of biology accessible to small-molecule intervention and reduce reliance on resource-intensive high-throughput screening. Towards this goal, the system prioritizes compounds for rapid experimental validation by coupling model-driven ranking with streamlined procurement from commercial catalogs. To improve developability at the point of selection, we introduce a suite of ADMET models for kinetic solubility (logS), lipophilicity (logD), and Caco-2 permeability. These models act as an early triage layer, systematically filtering out compounds with unfavorable physicochemical and absorption properties prior to synthesis or purchase. Across a panel of ten targets (most with no representation in the underlying affinity training data) we observe strong prospective performance on challenging systems. Functional actives or binders were identified for 6 of 10 targets, despite modest experimental budgets of 28-96 compounds per target. These results include successes on receptors and enzymes traditionally considered difficult for structure- or ligand-based approaches. Collectively, this work establishes a practical framework for low-throughput, cost-constrained discovery campaigns capable of delivering chemically tractable binders with favorable property profiles. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=124 SRC="FIGDIR/small/736485v1_fig1.gif" ALT="Figure 1"> View larger version (50K): org.highwire.dtl.DTLVardef@1d7ff06org.highwire.dtl.DTLVardef@1a7fc7aorg.highwire.dtl.DTLVardef@1b0dc98org.highwire.dtl.DTLVardef@62b978_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFigure 1:C_FLOATNO Overview of the prospective virtual-screening campaigns across all targets. For each target, the panel shows the predicted protein-ligand complex together with the number of compounds tested, the number of confirmed actives/binders, and the assays used for screening and follow-up. C_FIG

biochemistry↗

bioRxiv2026

BoltzProt-1: Towards Efficient De Novo Binder Design with Good Developability

Designing binders against novel protein targets remains a central challenge in computational drug discovery. Here we introduce BoltzProt-1, a pipeline for generating protein binders, including nanobodies, with improved hit rates and favorable developability properties. At its core lie a refined iteration of BoltzGens generative model and a novel protein-protein interaction prediction model, BoltzPPI. Employing BoltzPPI instead of BoltzGens standard structure-prediction confidence metrics to rank nanobody (VHH) designs increases the confirmed-binder hit rate from 3.3% to 8.0% across 10 novel targets. Assessed on 10 additional targets used in prior literature, the BoltzProt-1 pipeline obtains nanobody screening hits for 7 of 10 targets, surpassing the 6 of 10 previously reported by Chai-2. Finally, evaluating the developability of BoltzProt-1-designed nanobodies in terms of stability, aggregation, purity, polyspecificity and hydrophobicity reveals that 58% of its confirmed binders pass every criterion, exceeding both BoltzGen (40%) and clinical-stage VHH controls (21%). O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=104 SRC="FIGDIR/small/733997v1_ufig1.gif" ALT="Figure 1"> View larger version (39K): org.highwire.dtl.DTLVardef@d86de1org.highwire.dtl.DTLVardef@115da7eorg.highwire.dtl.DTLVardef@1bbb4e9org.highwire.dtl.DTLVardef@626a68_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2026

Quantitative Assessment of shRNA Loading and Delivery Efficiency of Engineered Extracellular Vesicles

RNA interference (RNAi) therapeutics enable selective silencing of disease-associated genes. Yet, their clinical application remains largely confined to the liver due to extrahepatic delivery constraints of current platforms such as GalNAc conjugates and lipid nanoparticles. Extracellular vesicles (EVs) offer an attractive alternative delivery strategy owing to their biocompatibility, ability to traverse biological barriers, and amenability to engineering. However, EV-mediated RNA delivery is limited by inefficient endogenous RNA loading and poor cytosolic release following uptake. Here, we establish a modular EV-based platform that addresses both challenges by integrating enhanced endogenous shRNA loading with fusogen-mediated cytosolic delivery. Using Argonaute 2 (AGO2)-assisted loading, we substantially increase shRNA copy numbers per vesicle (up to 3.7 copies/EV) and enable quantitative, molecule-resolved assessment of delivery potency. Engineered EVs achieve robust and reproducible shRNA-mediated gene silencing with picomolar IC50 values across multiple cell types and induce significant target knockdown in the mouse brain following intracerebral administration. Together, these findings demonstrate that coordinated engineering of shRNA loading and cytosolic release can overcome key limitations of EV-mediated small RNA delivery.

bioengineering↗

bioRxiv2025

BoltzGen: Toward Universal Binder Design

We introduce BoltzGen, an all-atom generative model for designing proteins and peptides across all modalities to bind a wide range of biomolecular targets. BoltzGen builds strong structural reasoning capabilities about target-binder interactions into its generative design process. This is achieved by unifying design and structure prediction, resulting in a single model that also reaches state-of-the-art folding performance. BoltzGens generation process can be controlled with a flexible design specification language over covalent bonds, structure constraints, binding sites, and more. We experimentally validate these capabilities in eight diverse design campaigns with functional and affinity readouts across 26 targets. In our experiments, binder modalities span from nanobodies to disulfide-bonded peptides, and targets from disordered proteins to small molecules. In particular, we identify nanobody binders for novel targets with low similarity to proteins with already known bound structures. We release model weights, data, and both inference and training code at: https://github.com/HannesStark/boltzgen.

bioengineering↗

bioRxiv2025

Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction

Accurately modeling biomolecular interactions is a central challenge in modern biology. While recent advances, such as AlphaFold3 and Boltz-1, have substantially improved our ability to predict biomolecular complex structures, these models still fall short in predicting binding affinity, a critical property underlying molecular function and therapeutic efficacy. Here, we present Boltz-2, a new structural biology foundation model that exhibits strong performance for both structure and affinity prediction. Boltz-2 introduces controllability features including experimental method conditioning, distance constraints, and multi-chain template integration for structure prediction, and is, to our knowledge, the first AI model to approach the performance of free-energy perturbation (FEP) methods in estimating small molecule-protein binding affinity. Crucially, it achieves strong correlation with experimental readouts on many benchmarks, while being at least 1000x more computationally efficient than FEP. By coupling Boltz-2 with a generative model for small molecules, we demonstrate an effective workflow to find diverse, synthesizable, high-affinity binders, as estimated by absolute FEP simulations on the TYK2 target. To foster broad adoption and further innovation at the intersection of machine learning and biology, we are releasing Boltz-2 weights, inference, and training code 1 under a permissive open license, providing a robust and extensible foundation for both academic and industrial research.

molecular biology↗

bioRxiv2025

Neuromorpha vorax: a previously unculturable cosmopolitan protist with an unexpectedly complex life-cycle belonging to Glissomonadida Clade-U/Group-TE

Soil protists are increasingly recognized as common members of complex communities that associate with plant root systems, though their contributions to these communities and to the plant host remain obscure. Members Clade-U/Group-TE, are cosmopolitan soil protists and are often among the most abundant protists associated with plant roots. Here we describe the isolation, culturing and characterization of Neuromorpha vorax, a member of the previously uncultured Clade-U/Group-TE branch of the order Glissomonadida. N. vorax grew readily when provided bacteriophage lysate as a food source. This allowed us to grow large numbers of the organisms from single-cell isolates and provided ideal conditions for following transitions from one morphology to another. N. vorax, like most Glissomonads, has a small, flagellated gliding form, but it also displays a wide range of other morphologies including: a crawling form; small and large trophozoites with multiple, long filopodia; small and large resting cysts; clusters of large dividing cells and it displayed cannibalistic feeding behaviors. Given the small size of most Glissomonads, it may be that other members of this important group, known from environmental surveys, but currently uncultured, might also be readily grown on bacteriophage lysates. In addition, given an abundant food source and clear viewing by microscopy, Glissomonads and other small protists may be found to have life cycles and behaviors that are more complex than is currently appreciated. ImportanceProtists from the Clade-U/Group-TE cluster of Glissomonads are wide-spread and abundant colonizers plant roots. Despite being known for over 30 years, they have remained uncultured. We show these protists can be easily cultured using an unusual food source, viral lysates of bacteria. This culturing method allows growth of high numbers of these organisms and reveal thatthey have an unexpectedly complex lifecycle that includes community feeding and cannibalism. Some other, currently unculturable, protists can perhaps be grown with these methods and many of these may also show unexpectedly complex lifecycles. The growth of eukaryotes on virus lysates raises the possibility that viruses in soil may directly contribute to the growth (and not just death) of eukaryotes in soil and root-associated communities.

microbiology↗

bioRxiv2024

Immortalization of mesenchymal stromal cells by hTERT does not affect the functional properties of secreted extracellular vesicles

Mesenchymal stromal cell-derived extracellular vesicles (MSC-EVs) have emerged as promising and safe therapeutic agents, however, donor heterogeneities, limited replicative life span and changes in the cellular phenotype throughout in vitro cultivation remain major hurdles for scalable EV production. For these reasons, this study aims to investigate the use of hTERT immortalized ( telomerized) MSCs as a potential source for efficient, standardized, reliable MSC-EVs production by comparing parental primary to their telomerized MSC counterparts. We observed that hTERT expression does not affect cell morphology or cellular doubling time, while ensuring unlimited, stable in vitro propagation. In addition, telomerized WJ-MSCs maintained the canonical expression profile of surface markers and the tri-lineage differentiation potential of their primary counterparts. In terms of EV characteristics, the immortalization by hTERT expression did not affect size, number, cargo composition or biological activity regarding anti-inflammatory, anti-fibrotic and wound healing properties in vitro. In summary, the use of hTERT to immortalize MSCs leads to the creation of cell lines that continuously produce MSC-EVs without altering any key functionalities of the cells or resulting EVs. This suggests that telomerization of human cells from single donors is a promising strategy for generating cell factories that can produce EVs in standardized conditions and at scale and with standardization.

cell biology↗

bioRxiv2024

Boltz-1: Democratizing Biomolecular Interaction Modeling

Understanding biomolecular interactions is fundamental to advancing fields like drug discovery and protein design. In this paper, we introduce BO_SCPLOWOLTZC_SCPLOW-1, an open-source deep learning model incorporating innovations in model architecture, speed optimization, and data processing achieving AO_SCPLOWLPHAC_SCPLOWFO_SCPLOWOLDC_SCPLOW3-level accuracy in predicting the 3D structures of biomolecular complexes. BO_SCPLOWOLTZC_SCPLOW-1 demonstrates a performance on-par with state-of-the-art commercial models on a range of diverse benchmarks, setting a new benchmark for commercially accessible tools in structural biology. Further, we push the boundary of capabilities of these models with BO_SCPLOWOLTZC_SCPLOWO_SCPCAP-C_SCPCAPO_SCPLOWSTEERINGC_SCPLOW, a new inference time steering technique that is able to fix hallucinations and non-physical predictions from the models. By releasing the training and inference code, model weights, datasets, and benchmarks under the MIT open license, we aim to foster global collaboration, accelerate discoveries, and provide a robust platform for advancing biomolecular modeling.

biophysics↗

bioRxiv2024

PINDER: The protein interaction dataset and evaluation resource

Protein-protein interactions (PPIs) are fundamental to understanding biological processes and play a key role in therapeutic advancements. As deep-learning docking methods for PPIs gain traction, benchmarking protocols and datasets tailored for effective training and evaluation of their generalization capabilities and performance across real-world scenarios become imperative. Aiming to overcome limitations of existing approaches, we introduce PINDER, a comprehensive annotated dataset that uses structural clustering to derive non-redundant interface-based data splits and includes holo (bound), apo (unbound), and computationally predicted structures. PINDER consists of 2,319,564 dimeric PPI systems (and up to 25 million augmented PPIs) and 1,955 high-quality test PPIs with interface data leakage removed. Additionally, PINDER provides a test subset with 180 dimers for comparison to AlphaFold-Multimer without any interface leakage with respect to its training set. Unsurprisingly, the PINDER benchmark reveals that the performance of existing docking models is highly overestimated when evaluated on leaky test sets. Most importantly, by retraining DiffDock-PP on PINDER interface-clustered splits, we show that interface cluster-based sampling of the training split, along with the diverse and less leaky validation split, leads to strong generalization improvements.

bioinformatics↗

bioRxiv2024

PLINDER: The protein-ligand interactions dataset and evaluation resource

Protein-ligand interactions (PLI) are foundational to small molecule drug design. With computational methods striving towards experimental accuracy, there is a critical demand for a well-curated and diverse PLI dataset. Existing datasets are often limited in size and diversity, and commonly used evaluation sets suffer from training information leakage, hindering the realistic assessment of method generalization capabilities. To address these shortcomings, we present PLIN-DER, the largest and most annotated dataset to date, comprising 449,383 PLI systems, each with over 500 annotations, similarity metrics at protein, pocket, interaction and ligand levels, and paired unbound (apo) and predicted structures. We propose an approach to generate training and evaluation splits that minimizes task-specific leakage and maximizes test set quality, and compare the resulting performance of DiffDock when retrained with different kinds of splits.

biochemistry↗

bioRxiv2024

Snorkel-tag Based Affinity Chromatography for Recombinant Extracellular Vesicle Purification

Extracellular vesicles (EVs) are lipid nanoparticles and play an important role in cell-cell communications, making them potential therapeutic agents and allowing to engineer for targeted drug delivery. The expanding applications of EVs in next generation medicine are still limited by existing tools for scaling standardized EV production, single EV tracing and analytics, and thus provide only a snapshot of tissue-specific EV cargo information. Here, we present CD81, an EV surface marker protein, genetically fused to series of tags with additional transmembrane domain to be displayed on the EV surface, which we term Snorkel-tag. This system enables to affinity purify EVs from complex matrices in a non-destructive form. In future applications, this strategy will allow generating transgenic animals to enable tracing and analyzing EVs, and their cargo in physiological and pathophysiological set-ups, and facilitate the development of EV based diagnostic tools in murine models which can be translated to humans.

bioengineering↗

bioRxiv2024

Cycling reduces the entropy of neuronal activity in the human adult cortex

Electroencephalogram (EEG) data is often analyzed from a Brain Complexity (BC) perspective, having successfully been applied to study the brain in both health and disease. In this study, we employed recurrence entropy to quantify BC associated with the neurophysiology of movement by comparing BC in both resting state and cycling movement. We measured EEG in 24 healthy adults, and placed the electrodes on occipital, parietal, temporal and frontal sites, on both the right and left sides. EEG measurements were performed for cycling and resting states and for eyes closed and open. We then computed recurrence entropy for the acquired EEG series. Our results show that open eyes show larger entropy compared to closed eyes; the entropy is also larger for resting state, compared to cycling state for all analyzed brain regions. The decrease in neuronal complexity measured by the recurrence entropy could explain the neural mechanisms involved in how the cycling movements suppress the freezing of gate in patients with Parkinsons disease due to the constant sensory feedback caused by cycling that is associated with entropy reduction.

neuroscience↗

bioRxiv2023

Novel endogenous engineering platform for robust loading and delivery of functional mRNA by extracellular vesicles

Messenger RNA (mRNA) has emerged as an attractive therapeutic molecule for a plethora of clinical applications. For in vivo functionality, mRNA therapeutics require encapsulation into effective, stable, and safe delivery systems to protect the cargo from degradation and reduce immunogenicity. Here, a bioengineering platform for efficient mRNA loading and functional delivery using bionormal nanoparticles, Extracellular Vesicles (EVs), is established by expressing a highly specific RNA-binding domain fused to CD63 in EV producer cells stably expressing the target mRNA. The additional combination with a fusogenic endosomal escape moiety, VSVg, enables functional mRNA delivery in vivo at doses substantially lower than currently used clinically with synthetic lipid-based nanoparticles. Importantly, the application of EVs loaded with effective cancer immunotherapy proves highly effective in an aggressive melanoma mouse model. This technology addresses substantial drawbacks currently associated with EV-based nucleic acid delivery systems and is a leap forward to clinical EV applications. Graphical Abstract(Figure created using BioRender) O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=189 SRC="FIGDIR/small/533081v3_ufig1.gif" ALT="Figure 1"> View larger version (56K): org.highwire.dtl.DTLVardef@140058eorg.highwire.dtl.DTLVardef@2ceaccorg.highwire.dtl.DTLVardef@eafd52org.highwire.dtl.DTLVardef@19d4019_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioengineering↗

bioRxiv2021

A dual sequence and culture-based survey of maize rhizosphere protists reveals dominant,plant-enriched, and culturable community members

Protists play important roles in shaping the microbial community of the rhizosphere. However, there is still a limited understanding of how plants shape the protist community, and how well protist isolate collections might represent rhizosphere protist composition and function in downstream studies. We sought to determine whether maize roots select for a distinct protist community in the field, and whether the common or dominant members of that community are readily culturable using standard protist isolation methods. We sequenced 18S and 16S rRNA genes from the rhizospheres of maize grown in two sites, and isolated 103 protists into culture from the same roots. While field site had the greatest effect, rhizospheres in both sites had distinct protist composition from the bulk soils, and certain taxa were enriched in both sites. Enriched taxa were correlated to bacterial abundance patterns. The isolated protists represented six supergroups, and the majority corresponded to taxa found in the sequencing survey. Twenty-six isolates matched eight of the 89 core rhizosphere taxa. This study demonstrates that maize roots select for a distinct protist community, but also illustrate the potential challenges in understanding the function of the dominant protist groups in the rhizosphere. Originality-Significance StatementThis is the first study comparing cultivation-dependent and independent methods for studying the protist community of plant roots, and the first untargeted analysis of the maize rhizospheres effect on protist communities. We show that maize in different sites select for distinct communities and overlapping enriched taxa, but that isolating the most important plant-associated protists may be a challenge for researchers.

microbiology↗

bioRxiv2020

Profiling of extracellular small RNAs highlights a strong bias towards non-vesicular secretion

Extracellular environment consists of a plethora of different molecules, including extracellular miRNA that can be secreted in association with extracellular vesicles (EVs) or soluble protein complexes (non-EVs). Yet, it is generally accepted that most of the biological activity is attributed to EV-associated miRNAs. The capability of EVs to transport cargoes has attracted much interest towards developing EVs as therapeutic short RNA carriers by using endogenous loading strategies for miRNA enrichment. Here, by overexpressing miRNA and shRNA sequences of interest in source cells and using size exclusion liquid chromatography (SEC) to separate the cellular secretome into EV and non-EV fractions, we saw that strikingly, <2% of all secreted overexpressed miRNA were found in association with EVs. To see whether the prominent non-EV miRNA secretion also holds true at the basal expression level of native miRNA transcripts, both fractions were further analysed by small RNA sequencing. This revealed a global correlation of EV and non-EV miRNA abundance to that of their parent cells and showed an enrichment only for miRNAs with a relatively low cellular expression level. Further quantification showed that similarly to the transient overexpression context, an outstanding 96.2-99.9% of total secreted miRNA at its basal level was secreted to the non-EV fraction. Yet, even though EVs contain only a fraction of secreted miRNAs, these molecules were found stable at 37{degrees}C in serum-containing environment, indicating that if sufficient miRNA loading to EVs is achieved, EVs can remain miRNA delivery-competent for a prolonged period of time. This study suggests that the passive endogenous EV loading strategy can be a relatively wasteful way of loading miRNA to EVs and active miRNA loading approaches are needed for developing advanced EV miRNA therapies in the future.

molecular biology↗

bioRxiv2020

Engineering of extracellular vesicles for display of protein biotherapeutics

Extracellular vesicles (EVs) have recently emerged as a highly promising cell-free bio-therapeutics. While a range of engineering strategies have been developed to functionalize the EV surface, current approaches fail to address the limitations associated with endogenous surface display, pertaining to the heterogeneous display of commonly used EV-loading moieties among different EV subpopulations. Here we present a novel engineering platform to display multiple protein therapeutics simultaneously on the EV surface. As proof-of-concept, we screened multiple endogenous display strategies for decorating the EV surface with cytokine binding domains derived from tumor necrosis factor receptor 1 (TNFR1) and interleukin 6 signal transducer (IL6ST), which can act as decoys for the pro-inflammatory cytokines TNF and IL6, respectively. Combining synthetic biology and systematic screening of loading moieties, resulted in a three-component system which increased the display and decoy activity of TNFR1 and IL6ST, respectively. Further, this system allowed for combinatorial functionalization of two different receptors on the same EV surface. These cytokine decoy EVs significantly ameliorated disease phenotypes in three different inflammatory mouse models for systemic inflammation, neuroinflammation, and intestinal inflammation. Importantly, significantly improved in vitro and in vivo efficacy of these engineered EVs was observed when compared directly to clinically approved biologics targeting the IL6 and TNF pathways.

bioengineering↗

bioRxiv2019

Impaired cholesterol metabolism in the mice model of cystic fibrosis. A vicious circle?

Patients with cystic fibrosis (CF) have low cholesterol absorption and, despite enhanced endogenous biosynthesis, low serum cholesterol. Herein, we investigated cholesterol metabolism in a murine CF model in comparison to wild type (WT) testing serum and liver surrogate biomarkers together with the hepatic expression of genes involved in cholesterol metabolism. CF mice display lower sterols absorption and increased endogenous biosynthesis. Subsequently, we evaluated the effects of a cholesterol-supplemented diet on cholesterol metabolism in CF and WT mice. The supplementation in WT mice determines biochemical changes similar to humans. Instead, CF mice with supplementation did not show significant changes, except for serum phytosterols (-50%), liver cholesterol (+35%) and TNF mRNA expression, that resulted 5-fold higher than in CF without supplementation. However, liver cholesterol in CF mice with supplementation resulted significantly lower compared to WT supplemented mice. This study shows that in CF mice there is a vicious circle in which the altered bile salts synthesis/secretion contribute to reduce cholesterol digestion/absorption. The consequence is the enhanced liver cholesterol biosynthesis that accumulates in the cell triggering inflammation.

biochemistry↗