bioRxiv : the preprint server for biology最新文献

Despite the recent breakthrough of AlphaFold (AF) in the field of protein sequence-to-structure prediction, modeling protein interfaces and predicting protein complex structures remains challenging, especially when there is a significant conformational change in one or both binding partners. Prior studies have demonstrated that AF-multimer (AFm) can predict accurate protein complexes in only up to 43% of cases.¹ In this work, we combine AlphaFold as a structural template generator with a physics-based replica exchange docking algorithm to better sample conformational changes. Using a curated collection of 254 available protein targets with both unbound and bound structures, we first demonstrate that AlphaFold confidence measures (pLDDT) can be repurposed for estimating protein flexibility and docking accuracy for multimers. We incorporate these metrics within our ReplicaDock 2.0 protocol²to complete a robust in-silico pipeline for accurate protein complex structure prediction. AlphaRED (AlphaFold-initiated Replica Exchange Docking) successfully docks failed AF predictions including 97 failure cases in Docking Benchmark Set 5.5. AlphaRED generates CAPRI acceptable-quality or better predictions for 63% of benchmark targets. Further, on a subset of antigen-antibody targets, which is challenging for AFm (20% success rate), AlphaRED demonstrates a success rate of 43%. This new strategy demonstrates the success possible by integrating deep-learning based architectures trained on evolutionary information with physics-based enhanced sampling. The pipeline is available at github.com/Graylab/AlphaRED.

Neural cell types have classically been characterized by their anatomy and electrophysiology. More recently, single-cell transcriptomics has enabled an increasingly fine genetically defined taxonomy of cortical cell types, but the link between the gene expression of individual cell types and their physiological and anatomical properties remains poorly understood. Here, we develop a hybrid modeling approach to bridge this gap. Our approach combines statistical and mechanistic models to predict cells' electrophysiological activity from their gene expression pattern. To this end, we fit biophysical Hodgkin-Huxley-based models for a wide variety of cortical cell types using simulation-based inference, while overcoming the challenge posed by the mismatch between the mathematical model and the data. Using multimodal Patch-seq data, we link the estimated model parameters to gene expression using an interpretable sparse linear regression model. Our approach recovers specific ion channel gene expressions as predictive of biophysical model parameters including ion channel densities, directly implicating their mechanistic role in determining neural firing.

Recent studies have highlighted the significance of the spindle midzone, the region between the segregating chromosomes, in ensuring proper chromosome segregation. By combining 3D electron tomography, cutting-edge light microscopy and a novel single cell in vitro essay allowing single molecule tracking, we have discovered a previously unknown role of the regulation of microtubule dynamics within the spindle midzone of C. elegans by the chromokinesin KLP-19, and its relevance for proper spindle function. Using Fluorescence recovery after photobleaching and a combination of second harmonic generation and two-photon fluorescence microscopy, we found that the length of the antiparallel microtubule overlap zone in the spindle midzone is constant throughout anaphase, and independent of cortical pulling forces as well as the presence of the microtubule bundling protein SPD-1. Further investigations of SPD-1 and KLP-19 in C. elegans, the homologs of PRC1 and KIF4a, suggest that KLP-19 regulates the overlap length and functions independently of SPD-1. Our data shows that KLP-19 plays an active role in regulating the length of microtubules within the midzone as well as the size of the antiparallel overlap region throughout mitosis. Depletion of KLP-19 in mitosis leads to an increase in microtubule length and thus microtubule-based interactions in the spindle midzone, which affects spindle dynamics and force transmission. Our data shows that by localizing KLP-19 to the spindle midzone in anaphase microtubule dynamics can be locally controlled allowing the formation of a functional midzone.

Rapid and high-fidelity phosphorylation of two serines (S32 and S36) of IκBα by a prototype Ser/Thr kinase IKK2 is critical for fruitful canonical NF-κB activation. Here, we report that IKK2 is a dual specificity Ser/Thr kinase that autophosphorylates itself at tyrosine residues in addition to its activation loop serines. Mutation of one such tyrosine, Y169, located in proximity to the active site, to phenylalanine, renders IKK2 inactive for phosphorylation of S32 of IκBα. Surprisingly, auto-phosphorylated IKK2 relayed phosphate group(s) to IκBα without ATP when ADP is present. We also observed that mutation of K44, an ATP-binding lysine conserved in all protein kinases, to methionine renders IKK2 inactive towards specific phosphorylation of S32 or S36 of IκBα, but not non-specific substrates. These observations highlight an unusual evolution of IKK2, in which autophosphorylation of tyrosine(s) in the activation loop and the invariant ATP-binding K44 residue define its signal-responsive substrate specificity ensuring the fidelity of NF-κB activation.

Double dipping is a well-known pitfall in single-cell and spatial transcriptomics data analysis: after a clustering algorithm finds clusters as putative cell types or spatial domains, statistical tests are applied to the same data to identify differentially expressed (DE) genes as potential cell-type or spatial-domain markers. Because the genes that contribute to clustering are inherently likely to be identified as DE genes, double dipping can result in false-positive cell-type or spatial-domain markers, especially when clusters are spurious, leading to ambiguously defined cell types or spatial domains. To address this challenge, we propose ClusterDE, a statistical method designed to identify post-clustering DE genes as reliable markers of cell types and spatial domains, while controlling the false discovery rate (FDR) regardless of clustering quality. The core of ClusterDE involves generating synthetic null data as an in silico negative control that contains only one cell type or spatial domain, allowing for the detection and removal of spurious discoveries caused by double dipping. We demonstrate that ClusterDE controls the FDR and identifies canonical cell-type and spatial-domain markers as top DE genes, distinguishing them from housekeeping genes. ClusterDE's ability to discover reliable markers, or the absence of such markers, can be used to determine whether two ambiguous clusters should be merged. Additionally, ClusterDE is compatible with state-of-the-art analysis pipelines like Seurat and Scanpy.

Occam's razor is the principle that, all else being equal, simpler explanations should be preferred over more complex ones. This principle is thought to guide human decision-making, but the nature of this guidance is not known. Here we used preregistered behavioral experiments to show that people tend to prefer the simpler of two alternative explanations for uncertain data. These preferences match predictions of formal theories of model selection that penalize excessive flexibility. These penalties emerge when considering not just the best explanation but the integral over all possible, relevant explanations. We further show that these simplicity preferences persist in humans, but not in certain artificial neural networks, even when they are maladaptive. Our results imply that principled notions of statistical model selection, including integrating over possible, latent causes to avoid overfitting to noisy observations, may play a central role in human decision-making.

Sensory cortical areas guide behavior by transforming stimulus-driven inputs into selective responses representing relevant features. A classic example is the representation of edge orientations in the visual cortex ^1-4 , where layer 4 (L4) neurons co-activated by an orientation provide feedforward inputs to specific functional modules in layer 2/3 (L2/3) that share strong recurrent connections ^5-7 . The aligned state of feedforward-recurrent interactions is critical for amplifying selective cortical responses ^8-12 , but how it develops remains unclear. Using simultaneous electrophysiology and calcium imaging in visually naïve animals we find that coactivity of L4 neurons and L2/3 modular responses elicited by oriented gratings lacks the tight relationship to orientation preference found in experienced animals. One factor that could contribute to the lack of functionally specific coactivity is high variability in naïve L4 neuron responses that decreases significantly following experience. But a computational model of feedforward-recurrent interaction suggests that high variability alone is insufficient to explain the naïve state and provides a biological signature of feedforward-recurrent misalignment that we confirm with whole-cell recordings: dynamic changes in orientation tuning of L2/3 subthreshold responses shortly after stimulus onset. In conclusion, we provide diverse evidence for a realignment of feedforward-recurrent interactions following experience that is critical for building reliable sensory representations with interlaminar temporal coherence.

Layer-dependent functional magnetic resonance imaging (fMRI) is a promising yet challenging approach for investigating layer-specific functional connectivity (FC). Achieving a brain-wide mapping of layer-specific FC requires several technical advancements, including sub-millimeter spatial resolution, sufficient temporal resolution, functional sensitivity, global brain coverage, and high spatial specificity. Although gradient echo (GE)-based echo planar imaging (EPI) is commonly used for rapid fMRI acquisition, it faces significant challenges due to the draining-vein contamination. In this study, we addressed these limitations by integrating velocity-nulling (VN) gradients into a GE-BOLD fMRI sequence to suppress vascular signals from the vessels with fast-flowing velocity. The extravascular contamination from pial veins was mitigated using a GE-EPI sequence at 3T rather than 7T, combined with phase regression methods. Additionally, we incorporated advanced techniques, including simultaneous multislice (SMS) acceleration and NOise Reduction with DIstribution Corrected principal component analysis (NORDIC PCA) denoising, to improve temporal resolution, spatial coverage, and signal sensitivity. This resulted in a VN fMRI sequence with 0.9-mm isotropic spatial resolution, a repetition time (TR) of 4 seconds, and brain-wide coverage. The VN gradient strength was determined based on results from a button-pressing task. Using resting-state data, we validated layer-specific FC through seed-based analyses, identifying distinct connectivity patterns in the superficial and deep layers of the primary motor cortex (M1), with significant inter-layer differences. Further analyses with a seed in the primary sensory cortex (S1) demonstrated the reliability of the method. Brain-wide layer-dependent FC analyses yielded results consistent with prior literature, reinforcing the efficacy of VN fMRI in resolving layer-specific functional connectivity. Given the widespread availability of 3T scanners, this technical advancement has the potential for significant impact across multiple domains of neuroscience research.

Although membrane proteins fold and function in a lipid bilayer constituting cell membranes, their structure and functionality can be recapitulated in diverse amphiphilic assemblies whose compositions deviate from native membranes. It remains unclear how various hydrophobic environments can stabilize membrane proteins and whether lipids play any role therein. Here, using the evolutionary unrelated α-helical and β-barrel membrane proteins of Escherichia coli , we find that the hydrophobic thickness and the strength of amphiphile- amphiphile packing are critical environmental determinants of membrane protein stability. Lipid solvation enhances stability by facilitating residue burial in the protein interior and strengthens the cooperative network by promoting the propagation of local structural perturbations. This study demonstrates that lipids not only modulate membrane proteins' stability but also their response to external stimuli.

How does functional modularity emerge in a cortex composed of repeats of a canonical local circuit? Focusing on distributed working memory, we show that a rigorous description of bifurcation in space describes the emergence of modularity. A connectome-based model of monkey cortex displays bifurcation in space during decision-making and working memory, demonstrating this new concept's generality. In a generative model and multi-regional cortex models of both macaque monkey and mouse, we found an inverted-V-shaped profile of neuronal timescales across the cortical hierarchy during working memory, providing an experimentally testable prediction of modularity. The cortex displays simultaneously many bifurcations in space, so that the corresponding modules could potentially subserve distinct internal mental processes. Therefore, a distributed process subserves the brain's functional specificity. We propose that bifurcation in space, resulting from connectivity and macroscopic gradients of neurobiological properties across the cortex, represents a fundamental principle for understanding the brain's modular organization.