Biology subjects

Lu, Y.-C.

Publications and source records attributed to Lu, Y.-C..

11 recordsLinked to original sources

bioRxiv2026

Long-term memory performance optimization via Neural network-based curve fitting in Drosophila

Long-term memory (LTM) formation typically requires extensive training or highly salient experiences, limiting learning efficiency. Operant conditioning is generally thought to produce stronger memory than classical conditioning because of its active learning component. Unexpectedly, however, laser-based social conditioning in Drosophila melanogaster revealed that while classical paradigms yielded lower short-term memory (STM) scores but higher LTM retention, operant paradigms exhibited higher STM scores followed by rapid LTM decay. To resolve this discrepancy, we employed the AI Complex Systems Response (AI-CSR) framework, which reconstructs high-dimensional learning landscapes from sparse sampling and predicts globally optimal training conditions. Following AI-CSR optimization, operant conditioning produced a twofold increase in LTM scores, yielding the strongest 24-hour social memory performance reported in flies to date and revealing the expected superiority of active learning, which had conversely shown poorer performance under standard training protocols. In contrast, AI-CSR did not further enhance classical conditioning performance but reduced training time by 50%. Single-cell RNA sequencing revealed expanded neuronal recruitment marked by the activation and inhibition of various gene combinations. Together, these findings link circuit-level reorganization with molecular programs underlying efficient long-term memory and demonstrate how AI-guided optimization can uncover latent learning capacity in biological systems.

animal behavior and cognition↗

bioRxiv2025

Characterization of mechanical tissue properties in post-mortem human brain using magnetic resonance elastography

Traumatic brain injury (TBI) is a serious health condition that can cause neurological dysfunction to varying degrees depending on the nature of the mechanical insult. In biomechanical studies of TBI under high loading conditions, post-mortem human subjects (PMHS) are often used since ethical concerns prohibit such experiments in living human subjects. Because PMHS brains undergo significant changes following death, it is important to understand the relationship between the mechanical properties of PMHS brain tissue and living tissue. In this study, we performed magnetic resonance elastography (MRE) on three PMHS specimens to estimate the material properties of the cadaveric brain, namely the storage modulus and the loss modulus, as well as the resulting shear stiffness and damping ratio. We also performed longitudinal MRE scans on one of the PMHS over the span of two months to investigate the evolution of tissue properties with post-mortem degradation. In comparison to in vivo subjects of age range 70-75 years, a substantially higher stiffness (mean: 5.96kPa) and lower damping ratio (mean: 0.09) were found in PMHS models. This study also revealed an initial increase in shear stiffness up to the seventh day post-mortem, followed by a steady decrease by the fifty-eighth day. However, the damping ratio displayed an opposite trend to that of shear stiffness. These changes were heterogeneous across brain regions. The collected measurements and analysis elucidate the changes in mechanical properties in post-mortem subjects, and can be used to build and validate computational models of TBI.

neuroscience↗

bioRxiv2025

Development and In Vivo Evaluation of a Novel Bioabsorbable Polylactic Acid Middle Ear Ventilation Tube

BackgroundOtitis media with effusion (OME) is a widespread condition causing hearing impairment, particularly in pediatric populations. Existing non-absorbable middle ear ventilation tubes frequently necessitate secondary surgical removal. Bioabsorbable polylactic acid (PLA) offers a promising alternative due to its inherent biocompatibility and tunable degradation characteristics. This study aimed to design, fabricate, and comprehensively evaluate a novel PLA middle ear ventilation tube. MethodsBioabsorbable PLA tubes were designed and fabricated based on commercial models. In vitro biocompatibility was assessed according to ISO 10993 guidelines. In vivo evaluations were performed in a guinea pig model, including otoscopic examinations, auditory brainstem response (ABR) measurements, micro-computed tomography (micro-CT) imaging, and histological analyses. ResultsThe PLA tubes were successfully designed and fabricated, exhibiting dimensions comparable to commercially available products. In vitro testing confirmed their biocompatibility. In vivo observations demonstrated that the PLA segments remained stable, with no significant inflammation detected. ABR measurements revealed no adverse impact on hearing function. Micro-CT imaging confirmed tube integrity and displayed initial signs of degradation over a 9-month period. Histological analyses indicated a favorable tissue response with minimal foreign body reaction. ConclusionThe developed PLA middle ear ventilation tube represents a highly promising alternative to conventional non-absorbable tubes. It demonstrates excellent biocompatibility, preserves auditory function, and exhibits a controlled degradation profile. This preclinical study provides strong support for further investigation and subsequent clinical trials to validate its safety and efficacy in human patients.

bioengineering↗

bioRxiv2024

scChat: A Large Language Model-Powered Co-Pilot for Contextualized Single-Cell RNA Sequencing Analysis

Single-cell RNA sequencing (scRNA-seq) has transformed biomedical research by enabling transcriptomic analysis at single-cell resolution. Yet, existing computational approaches remain primarily data-driven and lack the ability to integrate research context, limiting their interpretability and impact on hypothesis generation or experimental planning. We present scChat, a large language model (LLM)-powered co-pilot for contextualized scRNA-seq analysis. Unlike conventional pipelines restricted to tasks such as cell type annotation or enrichment analysis, scChat has an interactive, reasoning-based framework. It combines quantitative algorithms with retrieval-augmented generation and a multi-agent architecture to support hypothesis validation, mechanistic interpretation, and next-step experimental design. Through showcase and benchmarking studies, we demonstrate that scChat not only achieves high accuracy in cell type annotation but also provides biologically grounded explanations and contextual insights.

bioinformatics↗

bioRxiv2022

Cryo-EM reveals the structure and dynamics of a 723-residue malate synthase G

Determination of sub-100 kilodalton (kDa) structures by cryo-electron microscopy (EM) is a longstanding but not straightforward goal. Here, we present a 2.9-[A] cryo-EM structure of a 723-amino acid apo-form malate synthase G (MSG) from Escherichia coli. The cryo-EM structure of the 82-kDa MSG exhibits the same global folding as structures resolved by crystallography and nuclear magnetic resonance (NMR) spectroscopy, and the crystal and cryo-EM structures are indistinguishable. Analyses of MSG dynamics reveal consistent conformational flexibilities among the three structural approaches, most notably that the /{beta} domain exhibits heterogeneity. We observed that sidechains of F453, L454, M629, and E630 residues involved in hosting the cofactor acetyl-CoA and substrate rotate differently between the cryo-EM apo-form and complex crystal structures. Our work demonstrates that the cryo-EM technique can be used to determine structures and conformational heterogeneity of sub-100 kDa biomolecules to a quality as high as that obtained from crystallography and NMR spectroscopy.

biophysics↗

bioRxiv2022

Computer simulations reveal pathogenicity and inheritance modes of hearing loss-causing germinal variants

Variants in the gap junction beta-2 (GJB2) gene are the most common cause of hereditary hearing impairment. However, how GJB2 variants lead to local physicochemical and structural changes in the hexameric ion channels of connexin 26 (Cx26), resulting in hearing impairment, remains elusive. In the present study, using molecular dynamics (MD) simulations, we showed that detached inner-wall N-terminal "plugs" aggregated to reduce the channel ion flow in a highly prevalent V37I variant in humans. To examine the predictability of the computational platform, an artificial mutant, V37M, of which the effect was previously unknown in hearing loss, was created. Microsecond simulations showed that homomeric V37M Cx26 hemichannels had an abnormal affinity between the inner edge and N-termini to block the narrower side of the cone-shaped Cx26, while the most stable heteromeric channels did not. Consistent with these predictions, homozygous V37M transgenic mice exhibited apparent hearing loss, but not their heterozygous counterparts, indicating a recessive inheritance mode. Reduced channel conductivity was found in Gjb2V37M/V37M outer sulcus cells and Claudius cells but not in Gjb2WT/WT cells. We view that the current computational platform could serve as an assessment tool for the pathogenesis and inheritance of GJB2-related hearing impairments and other diseases caused by connexin dysfunction.

biophysics↗

bioRxiv2021

MSRCall: A Multi-scale Deep Neural Network to Basecall Oxford Nanopore Sequences

AO_SCPLOWBSTRACTC_SCPLOWMinION, a third-generation sequencer from Oxford Nanopore Technologies, is a portable device that can provide long nucleotide read data in real-time. It primarily aims to deduce the makeup of nucleotide sequences from the ionic current signals generated when passing DNA/RNA fragments through nanopores charged with a voltage difference. To determine the nucleotides from the measured signals, a translation process known as basecalling is required. However, compared to NGS basecallers, the calling accuracy of MinION still needs to be improved. In this work, a simple but powerful neural network architecture called MSRCall is proposed. MSRCall comprises a multi-scale structure, recurrent layers, a fusion block, and a CTC decoder. To better identify both short-range and long-range dependencies, the recurrent layer is redesigned to capture various time-scale features with a multi-scale structure. The results show that MSRCall outperforms other basecallers in terms of both read and consensus accuracies.

bioinformatics↗

bioRxiv2020

Expansion of Human Papillomavirus-Specific T Cells in Periphery and Cervix in a Therapeutic Vaccine Recipient Whose Cervical High-Grade Squamous Intraepithelial Lesion Regressed

Advances in high-throughput sequencing have revolutionized the manner with which we can study T cell responses. We describe a woman who received a human papillomavirus (HPV) therapeutic vaccine called PepCan, and experienced complete resolution of her cervical high-grade squamous intraepithelial lesion. By performing bulk T cell receptor (TCR) {beta} deep sequencing of peripheral blood mononuclear cells before and after 4 vaccinations, 70 putatively vaccine-specific clonotypes were identified for being significantly increased using a beta-binomial model. Interferon-{gamma} enzyme-linked immunospot assay identified significantly increased HPV-specific T cell response in the HPV 16 E6 91-115 region after 4 vaccinations (p=0.023). T cells with specificity to this region were sorted and analyzed using single-cell RNA-seq and TCR sequencing, HPV specificity in 60 of the 70 clonotypes identified to be vaccine-specific was demonstrated. TCR {beta} bulk sequencing of the cervical liquid-based cytology samples and cervical formalin-fixed paraffin-embedded samples before and after 4 vaccinations demonstrated the presence of these HPV-specific T cells in the cervix. Combining traditional and cutting-edge immunomonitoring techniques enabled us to demonstrate expansion of HPV-antigen specific T cells in the periphery and cervix.

immunology↗

bioRxiv2020

Toward the pathogenicity of the SLC26A4 p.C565Y variant using a genetically driven mouse model

Recessive variants of the SLC26A4 gene are a common cause of hearing impairment worldwide. In the past, cell lines and transgenic mice have been widely used to investigate the pathogenicity associated with the SLC26A4 variants. However, discrepancies in the pathogenicity between humans and cell lines or transgenic mice have been documented for some of the SLC26A4 variants. For instance, the p.C565Y variant, which has been reported to be pathogenic in humans, did not exhibit functional pathogenic consequences in cell lines. To address the pathogenicity of p.C565Y, we used a genotype-based approach in which we generated knock-in mice heterozygous (Slc26a4+/C565Y), homozygous (Slc26a4C565Y/C565Y), and compound heterozygous (Slc26a4919-2A>G/C565Y) for this variant. Subsequent phenotypic characterization revealed that mice segregating these genotypes demonstrated normal auditory and vestibular functions and normal inner ear morphology and expression of pendrin. These findings indicate that the p.C565Y variant is non-pathogenic for mice and that a single p.C565Y allele is sufficient to maintain normal inner ear physiology in mice. Our results highlight the differences in the pathogenicity associated with certain SLC26A4 variants between transgenic mice and humans, which should be taken into consideration while interpreting the results of animal studies for SLC26A4-related deafness.

genetics↗

bioRxiv2019

variant2literature: full text literature search for genetic variants

MotivationWhole genome sequencing (WGS) by next-generation sequencing produces millions of variants for an individual. The retrieval of biomedical literature for such a large number of genetic variants remains challenging, because in many cases the variants are only present in tables as images, or in the supplementary documents of which the file formats are diverse.\n\nResultsThe proposed tool named variant2literature from the TaiGenomics (Toolkits for AI genomics) resolves the problem by incorporating text recognition with image processing. In addition to the adoption of advanced image-based text retrieval, the recall rate of finding the literature containing the variants of interest is further improved by employing the skill of variant normalization. Different variant presentations are transformed into chromosome coordinates (standard VCF format) such that false negatives can be largely avoided. variant2literature is available in two ways. First, a web-based interface is provided to search all the literature in PMC Open Access Subset. Second, the command-line executable can be downloaded such that the users are free to search all the files in a specified directory locally.\n\nAvailabilityhttp://variant2literature.taigenomics.com/\n\nContactchienyuchen@ntu.edu.tw

genetics↗

bioRxiv2018

Single-cell microRNA/mRNA co-sequencing reveals non-genetic heterogeneity and novel regulatory mechanisms

Co-measurement of multiple omic profiles from the same single cells opens up the opportunity to decode molecular regulation that underlie intercellular heterogeneity in development and disease. Here, we present co-sequencing of microRNAs and mRNAs in the same single cells using a half-cell genomics approach. This method demonstrates good robustness (~95% success rate) and reproducibility (R2=0.93 for both miRNAs and mRNAs), and yields paired half-cell miRNA and mRNA profiles that could be independently validated. Linking the level of miRNAs to the expression of predicted target mRNAs across 19 single cells that are phenotypically identical, we observe that the predicted targets are significantly anti-correlated with the variation of abundantly expressed miRNAs, suggesting that miRNA expression variability alone may lead to non-genetic cell-to-cell heterogeneity. Genome-scale analysis of paired miRNA-mRNA co-profiles further allows us to derive and validate new regulatory relationships of cellular pathways controlling miRNA expression and variability.

systems biology↗