Biology subjects

Graves, A.

Publications and source records attributed to Graves, A..

8 recordsLinked to original sources

bioRxiv2026

SCREWx: A Screwless, Chronic, Recoverable, and Lightweight Neuropixels fixture for freely-moving rodents

High-density Neuropixels probes enable the study of large neural populations with single-cell and sub-millisecond resolution. While single-probe and acute head-fixed experiments have yielded critical scientific insights, understanding the neural mechanisms underlying many complex behaviors requires simultaneous multi-region recordings in freely moving, chronically implanted animals. Various probe fixtures have been developed to enable high-density recording, but existing designs impose critical limitations: their substantial weight restricts the maximum probe count that smaller animals can support, their bulky dimensions constrain the proximity of targeted brain regions, and their complex assembly risks damaging the probe during insertion and recovery. In this paper, we present a lightweight, fully 3D-printable, compact, and screwless fixture for chronic Neuropixels implants in freely moving rodents that features simple mechanisms for stable implantation and safe extraction. Our fixture design enables stable, high-yield single-unit recordings for months-long experiments, along with an 83% successful probe extraction rate. This fixture design provides a robust and accessible solution for long-term, multi-probe chronic Neuropixels recordings, increasing experimental throughput and enabling more complex experimental designs to investigate brain-wide neural dynamics.

neuroscience↗

bioRxiv2025

AbBFN2: A flexible antibody foundation model based on Bayesian Flow Networks

Antibody engineering is marked by diverse data and desiderata, making it a prime candidate for multi-objective design, but is commonly tackled as a series of sequential optimisation tasks. Here, we present AbBFN2, a generative antibody foundation model trained on paired antibody sequences as well as genetic and biophysical metadata. This is achieved using the Bayesian Flow Network paradigm, which allows unified modelling of diverse data sources and flexible conditional generation at inference time. By virtue of its rich set of features and architectural flexibility, AbBFN2 can be adapted to a number of tasks commonly tackled by individual models, consolidating traditional computational pipelines into a single step. We demonstrate the adaptability of AbBFN2 using sequence inpainting, humanisation, biophysical property optimisation, and conditional de novo library generation of antibodies with rare attributes as example tasks. By removing the need for task-specific training, we hope that AbBFN2 will accelerate machine learning-based antibody design and development workflows.

bioinformatics↗

bioRxiv2024

Protein Sequence Modelling with Bayesian Flow Networks

Exploring the vast and largely uncharted territory of amino acid sequences is crucial for understanding complex protein functions and the engineering of novel therapeutic proteins. Whilst generative machine learning has advanced protein sequence modelling, no existing approach is proficient for both unconditional and conditional generation. In this work, we propose that Bayesian Flow Networks (BFNs), a recently introduced framework for generative modelling, can address these challenges. We present ProtBFN, a 650M parameter model trained on protein sequences curated from UniProtKB, which generates natural-like, diverse, structurally coherent, and novel protein sequences, significantly outperforming leading autoregressive and discrete diffusion models. Further, we fine-tune ProtBFN on heavy chains from the Observed Antibody Space (OAS) to obtain an antibody-specific model, AbBFN, which we use to evaluate zero-shot conditional generation capabilities. AbBFN is found to be competitive with, or better than, antibody-specific BERT-style models, when applied to predicting individual framework or complimentary determining regions (CDR).

bioinformatics↗

bioRxiv2022

Image restoration of degraded time-lapse microscopy data mediated by infrared-imaging.

Time-lapse fluorescence microscopy is key to unraveling the processes underpinning biological development and function. However, living systems, by their nature, permit only a limited toolbox for interrogation. Consequently, following time-lapses, expended samples contain untapped information that is typically discarded. Herein we employ convolutional neural networks (CNNs) to augment the live imaging data using this complementary information. In particular, live, deep tissue imaging is limited by the spectral range of live-cell compatible probes/fluorescent proteins. We demonstrate that CNNs may be used to restore deep-tissue contrast in GFP-based time-lapse imaging using paired final-state datasets acquired using infrared dyes and improve information content accordingly. Surprisingly, the networks are remarkably robust over a wide range of developmental times. We employ said network to GFP time-lapse images captured during zebrafish and drosophila embryo/larval development and demonstrate live, deep tissue image contrast.

developmental biology↗

bioRxiv2022

Novel ERR pan-agonists ameliorate heart failure through boosting cardiac fatty acid metabolism and mitochondrial function

Cardiac metabolic dysfunction is a hallmark of heart failure. Estrogen related receptors ERR and ERR{gamma} are essential regulators for cardiac metabolism. Therefore, activation of ERR could be a potential therapeutic intervention for heart failure. However, no natural or synthetic ERR agonist is available to demonstrate their pharmacological effect in vivo. Using a structure-based design approach, we designed and synthesized two structurally distinct pan-ERR agonists, SLU-PP-332 (332) and SLU-PP-915 (915), which significantly improved ejection fraction and ameliorated fibrosis against pressure overload-induced heart failure without affecting cardiac hypertrophy. Mechanistically, a broad-spectrum of metabolic genes were transcriptionally activated by ERR agonists, particularly genes involved in fatty acid metabolism and mitochondrial function, which were mainly mediated by ERR{gamma}. Metabolomics analysis showed significant normalization of metabolic profiles in fatty acid/lipid and TCA/OXPHOS metabolites by 915 in the mouse heart with 6-week pressure overload. Autophagy was also induced by ERR agonists in cardiomycoyte. On the other hand, ERR agonism led to downregulation of cell cycle and development pathways, which was partially mediated by E2F1 in cardiomyocyte. In summary, ERR agonists maintain oxidative metabolism, which confers cardiac protection against pressure overload-induced heart failure in vivo. Our results provided direct pharmacological evidence supporting the further development of ERR agonists as novel heart failure therapeutics in vivo.

pharmacology and toxicology↗

bioRxiv2019

In vivo lineage conversion of vertebrate muscle into early endoderm-like cells

The extent to which differentiated cells, while remaining in their native microenvironment, can be reprogrammed to assume a different identity will reveal fundamental insight into cellular plasticity and impact regenerative medicine. To investigate in vivo cell lineage potential, we leveraged the zebrafish as a practical vertebrate platform to determine factors and mechanisms necessary to induce differentiated cells of one germ layer to adopt the lineage of another. We discovered that ectopic co-expression of Sox32 and Oct4 in several non-endoderm lineages, including skeletal muscle, can specifically trigger an early endoderm genetic program in a cell-autonomous manner. Gene expression, live imaging, and functional studies reveal that the endoderm-induced muscle cells lose muscle gene expression and morphology, while specifically gaining endoderm organogenesis markers, such as the pancreatic specification genes, hhex and ptf1a, via a mechanism resembling normal development. Endoderm induction by a pluripotent defective form of Oct4, endoderm markers appearing prior to loss of muscle cell morphology, a lack of dependence on cell division, and a lack of mesoderm, ectoderm, dedifferentiation, and pluripotency gene activation, together, suggests that reprogramming is endoderm specific and occurs via direct lineage conversion. Our work demonstrates that within a vertebrate animal, stably differentiated cells can be induced to directly adopt the identity of a completely unrelated cell lineage, while remaining in a distinct microenvironment, suggesting that differentiated cells in vivo may be more amenable to lineage conversion than previously appreciated. This discovery of possibly unlimited lineage potential of differentiated cells in vivo challenges our understanding of cell lineage restriction and may pave the way towards a vast new in vivo supply of replacement cells for degenerative diseases such as diabetes.

developmental biology↗

bioRxiv2019

The Pontine Nuclei are an Integrative Cortico-Cerebellar Link Critical for Dexterity

Skillful control of movement requires coordination between brain areas that are reciprocally connected through polysynaptic pathways, forming closed loops. A prominent loop in mammals runs between cerebral cortex and cerebellum, which individually contribute to skilled arm control. But how and why do these regions interact? Here, we studied the mouse cortico-cerebellar loop by optogenetically perturbing the pontine nuclei (PN), which receive direct cortical inputs and project only to cerebellum. PN stimulation during rest propagated into cerebellar cortex, but the effect of stimulation was transformed downstream into a wide range of patterns in the deep cerebellar nuclei (DCN) and reduced to transient excitation in motor cortex. PN stimulation in a cued reaching task altered arm kinematics and impaired performance. Cerebellar and cortical dynamics during movement were not dominated by PN stimulation, but altered in line with behavioral changes. These results suggest that the cortico-cerebellar loop fine-tunes motor commands during skilled reaching.

neuroscience↗

bioRxiv2018

Cortical Column and Whole Brain Imaging of Neural Circuits with Molecular Contrast and Nanoscale Resolution

Optical and electron microscopy have made tremendous inroads in understanding the complexity of the brain, but the former offers insufficient resolution to reveal subcellular details and the latter lacks the throughput and molecular contrast to visualize specific molecular constituents over mm-scale or larger dimensions. We combined expansion microscopy and lattice light sheet microscopy to image the nanoscale spatial relationships between proteins across the thickness of the mouse cortex or the entire Drosophila brain, including synaptic proteins at dendritic spines, myelination along axons, and presynaptic densities at dopaminergic neurons in every fly neuropil domain. The technology should enable statistically rich, large scale studies of neural development, sexual dimorphism, degree of stereotypy, and structural correlations to behavior or neural activity, all with molecular contrast.\n\nOne Sentence SummaryCombined expansion and lattice light sheet microscopy enables high speed, nanoscale molecular imaging of neural circuits over large volumes.

neuroscience↗