Biology subjects

Du, J.-L.

Publications and source records attributed to Du, J.-L..

4 recordsLinked to original sources

bioRxiv2026

NeuroPack: Advanced Learnable Packaging Toolkit Towards Ultra-High Compression Ratio for Volumetric Neuroimaging

High-throughput 3D microscopic imaging generates brain-volume datasets at terabyte to petabyte scale, posing significant storage challenges for large-scale neuroimaging data. In this study, we introduce NeuroPack, a hyperprior-driven 3D deep-learning-based packaging toolkit specifically designed for compressing 3D microscopic volumes at the mesoscale. NeuroPack encodes a compact hyperprior stream that conditions its entropy model on local voxel-wise variance, adaptively allocating bits to preserve mesoscale anatomical details for an optimal rate-distortion balance. To ensure reproducible deployment across heterogeneous microscopy datasets and computing environments, NeuroPack implements model integerization for the hyperprior side-information pathway, whose hyper-decoder predicts the Gaussian mean and scale parameters used by the arithmetic coder. By replacing platform-dependent floating-point hyper-decoding with deterministic fixed-point integer convolutions, NeuroPack produces bit-exact entropy models across CPUs, GPUs, local workstations, and cloud servers, thereby preventing decoding drift caused by floating-point rounding, fused-multiply-add ordering, or library-specific kernels. This integerized design preserves rate-distortion behavior while maintaining efficient encoding and decoding for large anisotropic 3D microscopy volumes. Our method enables high-fidelity microscopy reconstruction at 100- to 1000-fold compression across diverse modalities and species. Quantitatively, NeuroPack delivers the strongest overall rate-distortion performance in both BD-PSNR and BD-MSSSIM, with average relative gains of 73.4% and 50.8%, respectively, over the best competing codec. Importantly, NeuroPack preserves image fidelity required for downstream morphological reconstruction and analysis, as validated by simultaneous human annotation of original and decompressed confocal whole-brain zebrafish microscopy images. Together, NeuroPack provides a learning-based compression framework explicitly tailored for microscopy data, establishing a scalable storage solution for data-intensive neuroimaging and AI-empowered biological discovery.

bioengineering↗

bioRxiv2025

BrAVe: a unified framework for 3D interactive and integrative analysis of cross-modal multi-scale brain atlas data in diverse species

Understanding the brain requires integrating molecular, structural, and functional information across multiple scales. Despite the rapid expansion of multimodal, multiscale brain atlases in various species, integrative analysis remains challenging. Here, we present iBrAVE, an open-source platform for 3D interactive, integrative analysis of brain atlas datasets across modalities and scales within a common reference space. iBrAVE unifies heterogeneous datasets into five core representations, enabling source-independent analyses from insects to primates. At the single-modality level, it features spatially resolved quantification, as exemplified by single-neuron morphology analyses spanning geometry, topology, projection, clustering, and putative wiring. Across modalities, it links gene expression and neuronal activity with morphologically defined neurons and putative circuits via establishing spatial correspondence. Across scales, it enables bidirectional inference of neuronal identity and synaptic connectivity between light- and electron-microscopy reconstructions via neuronal morphology matching. Thus, iBrAVE provides a unified framework for integrative brain atlas analyses, facilitating data-driven discovery in neuroscience.

neuroscience↗

bioRxiv2025

Multiple-level organizational logic of the locus coeruleus-norepinephrine system in structure and function

The locus coeruleus (LC) regulates various neural processes through a limited number of norepinephrine (NE)-releasing neurons with widespread axon projections. However, how the LC is organized to exert brain-wide regulatory functions remain elusive. Utilizing larval zebrafish as a whole-brain-scale model, we reveal that morphologically heterogenous but physiologically comparable LC-NE neurons assemble complementarily to implement region-specific regulation of brain-wide neural dynamics. Individual LC-NE neurons exhibit diverse axon projections with ipsilateral bias and region preference, but share comparable electrophysiological properties, sensory responses and neural function-related genes expression. Moreover, population LC-NE neurons display bilaterally symmetrical yet regionally different axon projections and fire synchronously, causing regionally patterned NE release. Along with the regional comparability of NE receptor expression, these enable the LC to tune brain-wide neural dynamics in an inverted-U manner with region difference. Thus, our study provides a comprehensive understanding of the organizational logic of the LC system.

neuroscience↗

bioRxiv2020

Efficient replacement of long DNA fragments via non-homologous end joining at non-doding regions

Genomic DNA replacement for achieving sophisticated genetic manipulation is implemented currently through homogenous recombination/homology-dependent repair (HR/HDR). Here we report an efficient DNA fragment replacement method that is mediated by non-homologous end joining (NHEJ)-dependent DNA repair at two sites of CRISPR/Cas9-induced double-strand breaks at non-coding genomic regions flanking the exons of targeted genes. We demonstrated this method by generating three conditional alleles and two reporter lines of zebrafish. Functional assays of the conditional alleles proved that the genomic sequence between two inserted loxP sites was deleted by the Cre recombinase, and the phenotype after Cre-induced excision was comparable to previously reported mutants or morphants. Furthermore, combining double-fluorescence expression donor vectors, we showed that the efficiency of this NHEJ-mediated DNA replacement was around 3 times larger than that of HR/HDR-mediated approach. Our method provides a feasible strategy for genomic DNA replacement in zebrafish, which can be applicable for other organisms as well.

genetics↗