SEARCH · Search bioRxiv

Results for “biophysics”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 343 records · Page 19Linked to original sources

bioRxiv2022

MerMADE: Coupled biophysical, eco-evolutionary modelling for predicting population dynamics, movement and dispersal evolution in the marine environment

In order to understand patterns in species distributions, we need to understand the underlying mechanisms of dispersal, demography and evolutionary capability of these species. In the marine environment, few models combine these three key components likely due both to the computational challenges involved and the inherent challenges in data collection for parameterisation. To fill this gap, we have developed MerMADE, an individual-based, spatially explicit, eco-evolutionary coupled biophysical model for predicting population dynamics, dispersal and movement evolution in the marine environment (or aquatic environments in general). MerMADE combines dispersal in a 3D, hydrodynamically informed environment with population dynamics, demography and evolutionary functionality in order to investigate questions of connectivity, population persistence and evolution under environmental change and anthropogenic pressure. We illustrate its range of behavioural and physiological functionality using the lesser sandeel, Ammodytes marinus, as a case-study species in an invasion scenario. MerMADEs flexibility in species-specific parameterisation makes it a widely applicable, exciting tool in future sustainable management and conservation of aquatic species under environmental change.

ecology↗

bioRxiv2022

Conserved Biophysical Compatibility Among the Highly Variable Germline-Encoded Regions Shapes TCR-MHC Interactions

T cells are critically important components of the adaptive immune system primarily responsible for identifying and responding to pathogenic challenges. This recognition of pathogens is driven by the interaction between membrane-bound T cell receptors (TCRs) and antigenic peptides presented on major histocompatibility complex (MHC) molecules. The formation of the TCR-peptide-MHC complex (TCR-pMHC) involves interactions among germline-encoded and hypervariable amino acids. Germline-encoded and hypervariable regions can form contacts critical for complex formation, but only interactions between germline-encoded contacts are likely to be shared across many of all the possible productive TCR-pMHC complexes. Despite this, experimental investigation of these interactions have focused on only a small fraction of the possible interaction space. To address this, we analyzed every possible germline-encoded TCR-MHC contact in humans, thereby generating the first comprehensive characterization of these largely antigen-independent interactions. Our analysis shows that germline-encoded TCR-MHC interactions that are conserved at the sequence level are rare due to the high amino acid diversity of the TCR CDR1 and CDR2 loops, and that such conservation is unlikely to dominate the dynamic protein-protein binding interface. Instead, we find that binding properties such as the docking orientation are defined by regions of biophysical compatibility between these loops and the MHC surface.

immunology↗

bioRxiv2022

Prop3D: A flexible, Python-based platform for machine learning with protein structural properties and biophysical data

BackgroundMachine learning (ML) has a rich history in structural bioinformatics, and modern approaches, such as deep learning, are revolutionizing our knowledge of the subtle relationships between biomolecular sequence, structure, function, dynamics and evolution. As with any advance that rests upon statistical learning approaches, the recent progress in biomolecular sciences is enabled by the availability of vast volumes of sufficiently-variable data. To be useful, such data must be well-structured, machine-readable, intelligible and manipulable. These and related requirements pose challenges that become especially acute at the computational scales typical in ML. Furthermore, in structural bioinformatics such data generally relate to protein three-dimensional (3D) structures, which are inherently more complex than sequence-based data. A significant and recurring challenge concerns the creation of large, high-quality, openly-accessible datasets that can be used for specific training and benchmarking tasks in ML pipelines for predictive modeling projects, along with reproducible splits for training and testing. ResultsHere, we report Prop3D, a platform that allows for the creation, sharing and extensible reuse of libraries of protein domains, featurized with biophysical and evolutionary properties that can range from detailed, atomically-resolved physicochemical quantities (e.g., electrostatics) to coarser, residue-level features (e.g., phylogenetic conservation). As a community resource, we also supply a Prop3D-20sf protein dataset, obtained by applying our approach to CATH. We have developed and deployed the Prop3D framework, both in the cloud and on local HPC resources, to systematically and reproducibly create comprehensive datasets via the Highly Scalable Data Service (HSDS). Our datasets are freely accessible via a public HSDS instance, or they can be used with accompanying Python wrappers for popular ML frameworks. ConclusionProp3D and its associated Prop3D-20sf dataset can be of broad utility in at least three ways. Firstly, the Prop3D workflow code can be customized and deployed on various cloud-based compute platforms, with scalability achieved largely by saving the results to distributed HDF5 files via HSDS. Secondly, the linked Prop3D-20sf dataset provides a hand-crafted, already-featurized dataset of protein domains for 20 highly-populated CATH families; importantly, provision of this pre-computed resource can aid the more efficient development (and reproducible deployment) of ML pipelines. Thirdly, Prop3D-20sfs construction explicitly takes into account (in creating datasets and data-splits) the enigma of data leakage, stemming from the evolutionary relationships between proteins.

bioinformatics↗

bioRxiv2023

Biophysical characterization and modeling of SCN1A gain-of-function predicts interneuron hyperexcitability and a predisposition to network instability through homeostatic plasticity

SCN1A gain-of-function variants are associated with early onset developmental and epileptic encephalopathies (DEEs) that possess distinct clinical features compared to Dravet syndrome caused by SCN1A loss-of-function. However, it is unclear how SCN1A gain-of-function may predispose to cortical hyper-excitability and seizures. Here, we first report the clinical features of a patient carrying a de novo SCN1A variant (T162I) associated with neonatal-onset DEE, and then characterize the biophysical properties of T162I and three other SCN1A variants associated with neonatal-onset or early infantile DEE (I236V, P1345S, R1636Q). In voltage clamp experiments, three variants (T162I, P1345S and R1636Q) exhibited changes in activation and inactivation properties that enhanced window current, consistent with gain-of-function. Dynamic action potential clamp experiments utilising model neurons incorporating Nav1.1. channels supported a gain-of-function mechanism for all four variants. Here, the T162I, I236V, P1345S, and R1636Q variants exhibited higher peak firing rates relative to wild type and the T162I and R1636Q variants produced a hyperpolarized threshold and reduced neuronal rheobase. To explore the impact of these variants upon cortical excitability, we used a spiking network model containing an excitatory pyramidal cell (PC) and parvalbumin positive (PV) interneuron population. SCN1A gain-of-function was modeled by enhancing the excitability of PV interneurons and then incorporating three simple forms of homeostatic plasticity that restored pyramidal cell firing rates. We found that homeostatic plasticity mechanisms exerted differential impact upon network function, with changes to PV- to-PC and PC-to-PC synaptic strength predisposing to network instability. Overall, our findings support a role for SCN1A gain-of-function and inhibitory interneuron hyperexcitability in early onset DEE. We propose a mechanism through which homeostatic plasticity pathways can predispose to pathological excitatory activity and contribute to phenotypic variability in SCN1A disorders.

neuroscience↗

bioRxiv2023

Dissecting the Biophysical Origins of Activity-Stability Tradeoffs in D-amino Acid Oxidase with Enzyme Proximity-Seq

Understanding the complex relationships between enzyme sequence, folding stability and catalytic activity is crucial for applications in industry and biomedicine. However, current enzyme assay technologies are limited by an inability to simultaneously resolve both stability and activity phenotypes and to couple these to gene sequences at large scale. Here we developed Enzyme Proximity Sequencing (EP-Seq), a deep mutational scanning method that leverages peroxidase-mediated radical labeling with single cell fidelity to dissect the effects of thousands of mutations on stability and catalytic activity of oxidoreductase enzymes in a single experiment. We used EP-Seq to analyze how 6,399 missense mutations influence folding stability and catalytic activity in a D-amino acid oxidase (DAOx) from R.gracilis. The resulting datasets demonstrate activity-based constraints that limit folding stability during natural evolution, and identify hotspots distant from the active site as candidates for mutations that improve catalytic activity without sacrificing stability. EP-Seq can be extended to other enzyme classes and provides valuable insights into biophysical principles governing enzyme structure and function.

bioengineering↗

bioRxiv2023

A biophysical framework for double-drugging kinases

Orthosteric inhibition of kinases has been challenging due to the conserved active site architecture of kinases and emergence of resistance mutants. Simultaneous inhibition of distant orthosteric and allosteric sites, which we refer to as "double-drugging", has recently been shown to be effective in overcoming drug resistance. However, detailed biophysical characterization of the cooperative nature between orthosteric and allosteric modulators has not been undertaken. Here, we provide a quantitative framework for double-drugging of kinases employing isothermal titration calorimetry, Forster resonance energy transfer, coupled-enzyme assays, and X-ray crystallography. We discern positive and negative cooperativity for Aurora A kinase (AurA) and Abelson kinase (Abl) with different combinations of orthosteric and allosteric modulators. We find that a conformational equilibrium shift is the main principle governing this cooperative effect. Notably, for both kinases, we find a synergistic decrease of the required orthosteric and allosteric drug dosages when used in combination to inhibit kinase activities to clinically relevant inhibition levels. X-ray crystal structures of the doubledrugged kinase complexes reveal the molecular principles underlying the cooperative nature of double-drugging AurA and Abl with orthosteric and allosteric inhibitors. Finally, we observe the first fully-closed conformation of Abl when bound to a pair of positively cooperative orthosteric and allosteric modulators, shedding light onto the puzzling abnormality of previously solved closed Abl structures. Collectively, our data provide mechanistic and structural insights into rational design and evaluation of doubledrugging strategies.

biochemistry↗

bioRxiv2023

Controlling morpho-electrophysiological variability of neurons with detailed biophysical models

Variability is a universal feature among biological units such as neuronal cells as they enable a robust encoding of a high volume of information in neuronal circuits and prevent hyper synchronizations such as epileptic seizures. While most computational studies on electrophysiological variability in neuronal circuits were done with simplified neuron models, we instead focus on the variability of detailed biophysical models of neurons. With measures of experimental variability, we leverage a Markov chain Monte Carlo method to generate populations of electrical models able to reproduce the variability from sets of experimental recordings. By matching input resistances of soma and axon initial segments with the one of dendrites, we produce a compatible set of morphologies and electrical models that faithfully represent a given morpho-electrical type. We demonstrate our approach on layer 5 pyramidal cells with continuous adapting firing type and show that morphological variability is insufficient to reproduce electrical variability. Overall, this approach provides a strong statistical basis to create detailed models of neurons with controlled variability.

neuroscience↗

bioRxiv2023

Epistasis meets pleiotropy in shaping biophysical protein subspaces associated with antimicrobial resistance

Protein space is a rich analogy for genotype-phenotype maps, where amino acid sequence is organized into a high-dimensional space that highlights the connectivity between protein variants. It is a useful abstraction for understanding the process of evolution, and for efforts to engineer proteins towards desirable phenotypes. Few framings of protein space consider how higher-level protein phenotypes can be described in terms of their biophysical dimensions, nor do they rigorously interrogate how forces like epistasis--describing the nonlinear interaction between mutations and their phenotypic consequences--manifest across these dimensions. In this study, we deconstruct a low-dimensional protein space of a bacterial enzyme (dihydrofolate reductase; DHFR) into "subspaces" corresponding to a set of kinetic and thermodynamic traits [(kcat, KM, Ki, and Tm (melting temperature)]. We then examine how three mutations (eight alleles in total) display pleiotropy in their interactions across these subspaces. We extend this approach to examine protein spaces across three orthologous DHFR enzymes (Escherichia coli, Listeria grayi, and Chlamydia muridarum), adding a genotypic context dimension through which epistasis occurs across subspaces. In doing so, we reveal that protein space is a deceptively complex notion, and that the process of protein evolution and engineering should consider how interactions between amino acid substitutions manifest across different phenotypic subspaces.

systems biology↗

bioRxiv2023

Collecting-Gathering Biophysics of the Blackworm L. variegatus

Many organisms exhibit collecting and gathering behaviors as a foraging and survival method. Certain benthic macroinvertebrates are classified as collector-gatherers due to their collection of particulate matter as a food source, such as the aquatic oligochaete Lumbriculus variegatus (California blackworms). Blackworms demonstrate the ability to ingest organic and inorganic materials, including microplastics, but previous work has only qualitatively described their possible collecting behaviors for such materials. The mechanism through which blackworms consolidate discrete particles into a larger clumps remains unexplored quantitatively. By analyzing a group of blackworms in a large arena with an aqueous algae solution, we discover that their relative collecting efficiency is proportional to population size. Examining individual blackworms under a microscope reveals that both algae and microplastics physically adhere to the worms body due to external mucus secretions, which cause the materials to clump around the worm. We observe that this clumping reduces the worms exploration of its environment, potentially due to thigmotaxis. To validate the observed biophysical mechanisms, we create an active polymer model of a worm moving in a field of particulate debris with a short-range attractive force on its body to simulate its adhesive nature. We find that the attractive force increases gathering efficiency. This study offers insights into the mechanisms of collecting-gathering behavior, informing the design of robotic systems, as well as advancing our understanding the ecological impacts of microplastics on benthic invertebrates.

ecology↗

bioRxiv2023

Biophysical neural adaptation mechanisms enable deep learning models to capture dynamic retinal computation

Adaptation is a universal aspect of neural systems that changes circuit computations to match prevailing inputs. These changes facilitate efficient encoding of sensory inputs while avoiding saturation. Conventional artificial neural networks (ANNs) have limited adaptive capabilities, hindering their ability to reliably predict neural output under dynamic input conditions. Can embedding neural adaptive mechanisms in ANNs improve their performance? To answer this question, we develop a new deep learning model of the retina that incorporates the biophysics of photoreceptor adaptation at the front-end of conventional convolutional neural networks (CNNs). These conventional CNNs build on Deep Retina, a previously developed model of retinal ganglion cell (RGC) activity. CNNs that include this new photoreceptor layer outperform conventional CNN models at predicting primate and rat RGC responses to naturalistic stimuli that include dynamic local intensity changes and large changes in the ambient illumination. These improved predictions result directly from adaptation within the phototransduction cascade. This research underscores the potential of embedding models of neural adaptation in ANNs and using them to determine how neural circuits manage the complexities of encoding natural inputs that are dynamic and span a large range of light levels.

neuroscience↗

bioRxiv2023

A Biophysical Model of Visual Rivalry Links Cellular Mechanisms to Signatures of Conscious Perception

Contemporary models of perceptual awareness lack tractable neurobiological constraints. Inspired by recent cellular recordings in a mouse model of tactile threshold detection, we constructed a biophysical model of perceptual awareness that incorporated essential features of thalamocortical anatomy and cellular physiology. Our model reproduced, and mechanistically explains, the key in vivo neural and behavioural signatures of perceptual awareness in the mouse model, as well as the response to a set of causal perturbations. We generalised the same model (with identical parameters) to a more complex task - visual rivalry - and found that the same thalamic-mediated mechanism of perceptual awareness determined perceptual dominance. This led to the generation of a set of novel, and directly testable, electrophysiological predictions. Analyses of the model based on dynamical systems theory show that perceptual awareness in simulations of both threshold detection and visual rivalry arises from the emergent systems-level dynamics of thalamocortical loops.

neuroscience↗

bioRxiv2023

Biophysics of sodium channels during subthreshold depolarization in vitro and in silico

Voltage-gated sodium channels (VGSC) in the peripheral nervous system shape action potentials (AP) and thereby support the detection of sensory stimuli. Most of the nine mammalian VGSC subtypes are expressed in nociceptors, but predominantly, three are linked to several human pain syndromes: while Nav1.7 is suggested to be a (sub-)threshold channel, Nav1.8 is thought to support the fast AP upstroke. Nav1.9, as it produces large persistent currents, is attributed a role in determining the resting membrane potential. We characterized gating of Nav1.1-Nav1.3 and Nav1.5-Nav1.9 in manual patch clamp with focus on the AP subthreshold depolarization phase. Nav1.9 exhibited the most hyperpolarized activation while its fast inactivation resembled the depolarized inactivation of Nav1.8. For some VGSCs (e.g., Nav1.1 and Nav1.2), a positive correlation between ramp current and window current was detected. Using a modified Hodgkin-Huxley model which accounts for the time needed for inactivation to occur, we used the acquired data to simulate two nociceptive nerve fiber types (an A{delta}-and a mechano-insensitive C-nociceptor) containing VGSC conductances according to published human RNAseq data. Our simulations suggest that Nav1.9 is supporting both the AP upstroke and its shoulder. A reduced threshold for AP generation was induced by enhancing Nav1.7 conductivity or shifting its activation to more hyperpolarized potentials, as observed in Nav1.7-related pain disorders. Here, we provide a comprehensive, comparative functional characterization of VGSCs relevant in nociception and describe their gating with Hodgkin-Huxley-like models, which can serve as a tool to study their specific contributions to AP shape and sodium channel-related diseases. DisclaimerParts of this study were published as a preprint on bioRxiv: Koster, P.A., T. Stiehl, J. Tigerholm, A. Maxion, B. Namer, and A. Lampert. 2023. Biophysics of sodium channels during subthreshold depolarization in vitro and in silico. bioRxiv. doi.org/10.1101/2023.09.03.556095 (Preprint posted September 6, 2023) SummarySubthreshold gating of seven sodium channels (Nav1.1-3, Nav1.5-8) is determined by manual patch clamp and, together with Nav1.9, integrated into a computer model of an A{delta}-and a mechano-insensitive nociceptor (CMi). Simulations reveal contribution of Nav1.9 to the action potential upstroke and shoulder and prove useful for Nav1.7-related disease modelling.

physiology↗

bioRxiv2023

Biophysically Interpretable Inference of Cell Types from Multimodal Sequencing Data

Multimodal, single-cell genomics technologies enable simultaneous capture of multiple facets of DNA and RNA processing in the cell. This creates opportunities for transcriptome-wide, mechanistic studies of cellular processing in heterogeneous cell types, with applications ranging from inferring kinetic differences between cells, to the role of stochasticity in driving heterogeneity. However, current methods for determining cell types or clusters present in multimodal data often rely on ad hoc or independent treatment of modalities, and assumptions ignoring inherent properties of the count data. To enable interpretable and consistent cell cluster determination from multimodal data, we present meK-Means (mechanistic K-Means) which integrates modalities and learns underlying, shared biophysical states through a unifying model of transcription. In particular, we demonstrate how meK-Means can be used to cluster cells from unspliced and spliced mRNA count modalities. By utilizing the causal, physical relationships underlying these modalities, we identify shared transcriptional kinetics across cells, which induce the observed gene expression profiles, and provide an alternative definition for clusters through the governing parameters of cellular processes.

genomics↗

bioRxiv2023

Evidence for kilometer-scale biophysical features at the Gulf Stream front

Understanding the interplay of ocean physics and biology, particularly at the submesoscale and below (<30km), is an ongoing challenge in oceanography. While poorly constrained, these scales may be of critical importance for understanding how changing ocean dynamics will impact marine ecosystems. Fronts in the ocean, regions where two disparate water masses meet and isopycnals become tilted towards vertical, are considered hotspots for biophysical interaction, but there is limited observational evidence at the appropriate scales to assess their importance. Western boundary currents like the Gulf Stream are of particular interest as these dynamic physical regions are thought to influence both productivity and composition of primary producers; however, how exactly this plays out, and at what scales, is not well known. Using satellite data and two years of detailed in situ observations across the Gulf Stream front near Cape Hatteras, North Carolina, U.S.A., we investigate how submesoscale frontal dynamics could affect biological communities associated with frontal regions and generate hotspots of productivity and export. In this analysis we assess the seasonality and phenology of the region, generalize the kilometer-scale structure of the front, and analyze 69 transects to assess two physical processes of potential biogeochemical importance: cold shelf filament subduction and high salinity Sargasso Sea obduction. We link these processes observationally to the meander phase of the Gulf Stream and discuss how cold filament subduction could be exporting carbon and how obduction of high salinity water from depth often leads to high chlorophyll-a. Finally, we report on phytoplankton community composition in each of these features and integrate these new observations into our understanding of frontal submesoscale dynamics. Plain Language SummaryPhytoplankton move with large currents and are stirred by eddies with diameters ranging from 100s of kilometers down to the meter scale. Their growth is impacted by physical factors like light and temperature and also chemical and biological factors like nutrient availability, and their accumulation is also impacted by top down controls (zooplankton grazing, viral lysis) and competition with other phytoplankton. This interplay of physics and biology in determining the biomass and composition of phytoplankton communities is poorly understood and is key to understanding marine ecosystem resilience and structure in a changing ocean. In this work we investigated the impact of physics and biology on phytoplankton across scales focusing on the Gulf Stream front. Fronts in the ocean are where lines of equal density go from being horizontal to having a vertical tile, and because of this can enable nutrients and plankton to move from depth to the surface and vice versa. The objective of this work is to understand how physics might drive important changes in phytoplankton biomass and composition in the Gulf Stream front, which is amongst the sharpest gradients in temperature, density, and current speed in the global ocean. We find two frequent processes at the front, the apparent subduction of cold filaments down along the edge of the Gulf Stream, associated with meander troughs, and obduction of high salinity Sargasso Sea water into the front linked to meander crests. While ephemeral, these processes are frequent and could have a large impact on local phytoplankton biomass, phytoplankton composition, and the export of organic matter to depth. Key PointsO_LIThe frontal zone between the Gulf Stream and the shelf has an interface water mass which appears to have different origins with a range of biogeochemical impacts. C_LIO_LIMeanders appear to largely control the frontal interface: troughs lead to subduction of shelf filaments and crests lead to obduction of high salinity water. C_LIO_LIThese two processes are common at the front and could lead to ephemeral kilometer- scale export and productivity. C_LI

ecology↗

bioRxiv2023

Brain-derived and in vitro-seeded alpha-synuclein fibrils exhibit distinct biophysical profiles

The alpha-synuclein (Syn) seeding amplification assay (SAA) that allows the generation of disease-specific in vitro seeded fibrils (SAA fibrils) is used as a research tool to study the connection between the structure of Syn fibrils, cellular seeding/spreading, and the clinico-pathological manifestations of different synucleinopathies. However, structural differences between human brain-derived and SAA Syn fibrils have been recently highlighted. Here, we characterize biophysical properties of the human brain-derived Syn fibrils from the brains of patients with Parkinsons disease with and without dementia (PD, PDD), dementia with Lewy bodies (DLB), multiple system atrophy (MSA) and compare them to the model SAA fibrils. We report that the brain-derived Syn fibrils show distinct biochemical profiles, which were not replicated in the corresponding SAA fibrils. Furthermore, the brain-derived Syn fibrils from all synucleinopathies displayed a mixture of- straight and twisted microscopic structures. However, the PD, PDD, and DLB SAA fibrils had a straight structure, whereas MSA SAA fibrils showed a twisted structure. Finally, the brain-derived Syn fibrils from all four synucleinopathies were phosphorylated (S129). Interestingly, phosphorylated Syn were carried over to the PDD and DLB SAA fibrils. Our findings demonstrate the limitation of the SAA fibrils modelling the brain-derived Syn fibrils and pay attention to the necessity of deepening the understanding of the SAA fibrillation methodology.

neuroscience↗

bioRxiv2023

Setting a trajectory for CO2 emission reduction in academic research: a case study of a French biophysics laboratory

Climate change is a scientifically proven phenomenon caused by anthropic activities, which requires urgent and significant reductions in greenhouse gas emissions. Despite the increasing vocalization of scientists advocating for political action, the issue of the environmental impact of academic research has been neglected for some time. Now, field-dependent initiatives have emerged, such as the non-profit organization My Green Lab, which delivers green certifications to biology and chemistry labs, and institute-dependent programs, such as the Max Planck Sustainability Network. In France, an independent collective was founded in 2019 to address the environmental footprint of academic research following the COP 15 Paris Agreement: Labos 1Point5. Building on their resources and methodology, we have quantified the overall carbon footprint of our biophysics laboratory, considering energy consumption, purchases and travel, for the year 2021. We investigate how this footprint would decrease by 2030 following systemic changes (change in the energy mix, improvements from suppliers), and we propose scenarios based on additional voluntary initiatives to reach a final reduction of -50% compared to the 2021 baseline, following IPCC targets. We have now formed a group of more than 20 colleagues to achieve this goal, emphasizing the importance of collective action. Finally, we provide advice based on our own experience to assist others in addressing the environmental impact of academic research in their respective laboratories.

scientific communication and education↗

bioRxiv2023

Multitask Learning of Biophysically-Detailed Neuron Models

The human brain operates at multiple levels, from molecules to circuits, and understanding these complex processes requires integrated research efforts. Simulating biophysically-detailed neuron models is a computationally expensive but effective method for studying local neural circuits. Recent innovations have shown that artificial neural networks (ANNs) can accurately predict the behaviour of these detailed models in terms of spikes, electrical potentials, and optical readouts. While these methods have the potential to accelerate large network simulations by several orders of magnitude compared to conventional differential equation based modelling, they currently only predict voltage outputs for the soma or a select few neuron compartments. Our novel approach, based on enhanced state-of-the-art architectures for multitask learning (MTL), allows for the simultaneous prediction of membrane potentials in each compartment of a neuron model, at a speed of up to two orders of magnitude faster than classical simulation methods. By predicting all membrane potentials together, our approach not only allows for comparison of model output with a wider range of experimental recordings (patch-electrode, voltage-sensitive dye imaging), it also provides the first stepping stone towards predicting local field potentials (LFPs), electroencephalogram (EEG) signals, and magnetoencephalography (MEG) signals from ANN-based simulations. It further presents a challenging benchmark for MTL architectures due to the large amount of data involved, the presence of correlations between neighbouring compartments, and the non-Gaussian distribution of membrane potentials.

neuroscience↗

bioRxiv2023

ElectroPhysiomeGAN: Generation of Biophysical Neuron Model Parameters from Recorded Electrophysiological Responses

Recent advances in connectomics, biophysics, and neuronal electrophysiology warrant modeling of neurons with further details in both network interaction and cellular dynamics. Such models may be referred to as ElectroPhysiome, as they incorporate the connectome and individual neuron electrophysiology to simulate neuronal activities. The nervous system of C. elegans is considered a viable framework for such ElectroPhysiome studies due to advances in connectomics of its somatic nervous system and electrophysiological recordings of neuron responses. In order to achieve a simulated ElectroPhysiome, the set of parameters involved in modeling individual neurons need to be estimated from electrophysiological recordings. Here, we address this challenge by developing a deep generative estimation method called ElectroPhysiomeGAN (EP-GAN), which once trained, can instantly generate parameters associated with the Hodgkin-Huxley neuron model (HH-model) for multiple neurons with graded potential response. The method combines Generative Adversarial Network (GAN) architecture with Recurrent Neural Network (RNN) Encoder and can generate an extensive number of parameters (>170) given the neurons membrane potential responses and steady-state current profiles. We validate our method by estimating HH-model parameters for 200 simulated neurons with graded membrane potential followed by 9 experimentally recorded neurons (where 6 of them newly recorded) in the nervous system of C. elegans. Comparison of EP-GAN with existing estimation methods shows EP-GAN advantage in the accuracy of estimated parameters and inference speed for both small and large number of parameters being inferred. In addition, the architecture of EP-GAN permits input with arbitrary clamping protocols, allowing inference of parameters even when partial membrane potential and steady-state currents profile are given as inputs. EP-GAN is designed to leverage the generative capability of GAN to align with the dynamical structure of HH-model, and thus able to achieve such performance.

neuroscience↗