Biology subjects

Parker, K.

Publications and source records attributed to Parker, K..

10 recordsLinked to original sources

bioRxiv2026

Top Model Decision Tree: Selecting Segmentation Models for Reliable Quantitative Analysis in Low- and Ultralow-Dose CryoEM

Deep learning neural networks provide a powerful approach for segmenting low-contrast cryogenic electron microscopy (cryoEM) images. However, model performance can vary significantly across imaging conditions and may hinder downstream quantitative analyses. Here, we present a structured evaluation workflow to systematically screen segmentation models based on performance, inference speed, robustness across imaging conditions, and reliability of downstream quantitative measurements. Using the Bacterial Cell Envelope Thickness Tool (BCET) as a test case, we evaluate multiple architectures (YOLOv11, YOLO26, U-Net, Detectron2, and SAM3) under low-dose and ultralow-dose cryoEM conditions. While several models achieve high metrics, model choice strongly influences downstream measurements of envelope thickness. Models optimized for high F1-scores may produce unreliable segmentation masks from object crowding, interpolation artifacts or imaging conditions. Our results reveal distinct trade-offs between performance, speed, and robustness amongst models. YOLOv11 provides the highest fidelity membrane segmentation for quantitative measurements and the Meta-based model SAM3 offers improved robustness under ultralow-dose conditions with competitive inference performance. This work provides practical guidance for model selection in cryoEM workflows, emphasizing that optimal choice depends on experimental priorities and downstream analysis requirements rather than metrics alone. These findings are broadly relevant to cryoEM workflows as AI-based analysis expands beyond the biological sciences. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=142 SRC="FIGDIR/small/730486v1_ufig1.gif" ALT="Figure 1"> View larger version (33K): org.highwire.dtl.DTLVardef@f29df4org.highwire.dtl.DTLVardef@601d6eorg.highwire.dtl.DTLVardef@2c5023org.highwire.dtl.DTLVardef@1413f76_HPS_FORMAT_FIGEXP M_FIG C_FIG

microbiology↗

bioRxiv2026

Lateral Cerebellar Nucleus Lesions Disrupt Rule-Based Category Learning and Cognitive Flexibility in Rats

Cerebellar communication with the prefrontal cortex (PFC) may play a significant role in cognitive functions. Our previous studies found that rule-based (RB) category learning depends on the PFC in humans and rats. The PFC is also crucial for behavioral flexibility following rule-changes in various tasks. Very little is known regarding the role of the cerebellum in RB category learning. The current study was designed to determine whether the cerebellum plays a role in RB category learning, and in categorization following a rule switch. Female and male rats were given bilateral lesions of the lateral cerebellar nuclei (LCN) or a control surgery and trained on an RB category learning task followed by a category rule switch. A subset of rats was trained on a control discrimination task with the same trial procedures as the categorization task. Rats with LCN lesions took significantly longer to learn both the first and second category rules but were not impaired on the control task. Computational modeling revealed less task engagement and increased switching between engaged and non-engaged states in the LCN lesion group. Several measures of task performance indicated that the category learning deficit was not caused by a motor impairment, response bias, or an inability to discriminate the stimuli. The category learning deficits with LCN lesions were related to reduced accuracy of stimulus classification, an inability to maintain task engagement, and loss of flexibility. The results show, for the first time, that the cerebellum plays a crucial role in category learning and category rule-switching.

neuroscience↗

bioRxiv2025

Leveraging CryoEM and AI-Driven Morphological Feature Analysis for Insights on Bacterial Structures

Bacteria adapt by undergoing dynamic structural changes in response to environmental cues, which are often indicative of deeper phenotypic shifts in physiology and behavior. Understanding these changes across length scales is crucial for elucidating bacterial lifecycles, informing antifouling surface design, enhancing pathogen detection, and advancing renewable energy applications. Cryogenic electron microscopy (cryoEM) enables high-resolution imaging of bacterial structures in hydrated, biologically relevant conditions. However, current quantitative analysis strategies that extract structural information from bacterial samples remain labor-intensive. This work presents an AI-driven segmentation workflow tailored to low-dose cryoEM datasets to rapidly analyze bacterial ultrastructural features from Pantoea sp. YR343, a Gram-negative bacterium isolated from the rhizosphere of Populus deltoides that forms robust biofilms along plant roots. YOLOv11 image segmentation quantifies inner and outer membrane thickness and flagella length, enabling automated analysis of bacterial ultrastructure. The workflow reliably distinguishes membranes from carbon edges of the TEM grid and contaminant crystalline ice while matching manual measurements and increasing throughput. Flagella detection routines additionally quantify nearest-neighbor proximity between bacterial envelopes and flagella. A field-of-view module further detects bacteria at low magnification for rapid screening. Together, these tools provide a scalable and automated framework for high-throughput, quantitative analysis of bacterial ultrastructure in cryoEM data. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=39 SRC="FIGDIR/small/692658v1_ufig1.gif" ALT="Figure 1"> View larger version (21K): org.highwire.dtl.DTLVardef@18eac2org.highwire.dtl.DTLVardef@1dc67a1org.highwire.dtl.DTLVardef@1179499org.highwire.dtl.DTLVardef@11d08f8_HPS_FORMAT_FIGEXP M_FIG AI-based tools enable rapid characterization of bacterial ultrastructure in low-dose cryoEM. The envelope thickness tool quantifies membrane thickness and anisotropy. The flagella module analyzes filament morphology and detects cell-flagella contacts. The field-of-view (FOV) module C_FIG

microbiology↗

bioRxiv2023

16p11.2 Microduplication is Associated with Lobule-Specific Abnormalities in Cerebellar Structure and Function

The 16p11.2 microduplication (16p11.2dp/+) is associated with several neuropsychiatric disorders including schizophrenia, autism spectrum disorder, bipolar disorder, intellectual disability, and attention deficit/hyperactivity disorder (ADHD). Cerebellar abnormalities have been increasingly implicated in these neuropsychiatric disorders, including those conferred by 16p11.2 microduplication. In 16p11.2dp/+ mouse models, the cerebellum is a site of transcriptional dysregulation, and cerebellar microcephaly has been reported in humans with 16p11.2 microduplication. Despite mounting evidence indicating a role for the cerebellum in neuropsychiatric disorders associated with this CNV, cerebellar cellular structure and cerebellar-dependent behavior in mice with 16p11.2 microduplication remain uncharacterized. To address this, we histologically labeled Purkinje cells (PCs) and molecular layer interneurons (MLIs) in a mouse model of 16p11.2 microduplication. We did not find any structural differences in cerebellar lobule IV/V, nor did we observe impairments in gait or motor coordination, behaviors that are associated with lobule IV/V. In contrast, we discovered a significant increase in calbindin/parvalbumin-positive PCs mislocalized to the granule layer of cerebellar lobule VI in 16p11.2dp/+ mice compared to wild-type (WT) littermates. Additionally, we found a significant decrease in parvalbumin-positive MLIs without a decrease in total DAPI-positive cell counts in lobule VI of 16p11.2dp/+ mice compared to WT littermates. Cerebellar lobule VI is associated with delay eyeblink conditioning, and 16p11.2dp/+ mice are impaired in cerebellum-dependent associative learning on this task. Specifically, 16p11.2dp/+ mice showed deficits in both conditioned response (CR) percentage and CR onset latency relative to WT mice. These results suggest that lobule VI-specific alterations to PC localization and MLI parvalbumin expression in 16p11.2dp/+ mice impair both cerebellar learning and adaptive timing of cerebellar-driven, conditioned responses. Thus, we have identified novel structural and functional alterations in the cerebellum that are associated with 16p11.2 microduplication. Importantly, individuals with schizophrenia and ADHD also show CR acquisition deficits in delay eyeblink conditioning. Together, these data suggest that the behavioral impairments in 16p11.2dp/+ mice resemble impairments seen in neuropsychiatric disorders linked to 16p11.2 microduplication in humans. Further investigation of cerebellar cortex neurons in 16p11.2dp/+ mice may provide insights into the pathogenesis of neuropsychiatric disorders linked to this copy number variant.

neuroscience↗

bioRxiv2023

Enhancing Breast Ultrasound Segmentation through Fine-tuning and Optimization Techniques: Sharp Attention UNet

Segmentation of breast ultrasound images is a crucial and challenging task in computer-aided diagnosis systems. Accurately segmenting masses in benign and malignant cases and identifying regions with no mass is a primary objective in breast ultrasound image segmentation. Deep learning (DL) has emerged as a powerful tool in medical image segmentation, revolutionizing how medical professionals analyze and interpret complex imaging data. The UNet architecture is a highly regarded and widely used DL model in medical image segmentation. Its distinctive architectural design and exceptional performance have made it a popular choice among researchers in the medical image segmentation field. With the increase in data and model complexity, optimization and fine-tuning models play a vital and more challenging role than before. This paper presents a comparative study evaluating the effect of image preprocessing and different optimization techniques and the importance of fine-tuning different UNet segmentation models for breast ultrasound images. Optimization and fine-tuning techniques have been applied to enhance the performance of UNet, Sharp UNet, and Attention UNet. Building upon this progress, we designed a novel approach by combining Sharp UNet and Attention UNet, known as Sharp Attention UNet. Our analysis yielded the following quantitative evaluation metrics for the Sharp Attention UNet: the dice coefficient, specificity, sensitivity, and F1 score obtained values of 0.9283, 0.9936, 0.9426, and 0.9412, respectively. In addition, McNemars statistical test was applied to assess significant differences between the approaches. Across a number of measures, our proposed model outperforms the earlier designed models and points towards improved breast lesion segmentation algorithms.

bioengineering↗

bioRxiv2021

TLR2 on CD4+ and CD8+ T cells promotes late control of Mycobacterium tuberculosis infection

Although a role for TLR2 on T cells has been indicated in prior studies, in vivo stimulation of TLR2 on T cells by Mtb and its impact on Mtb infection has not been tested. Furthermore, it is not known if the enhanced susceptibility to Mtb of Tlr2 gene knockout (ko) mice is due to its role in macrophages, on T cells or both. To address TLR2 on T cells, we generated Tlr2fl/flxCd4cre/cre mice, which lack expression of TLR2 on both CD4 and CD8 T cells, to study the in vivo role of TLR2 on T cells after aerosol infection with virulent Mtb. Deletion of TLR2 in CD4+ and CD8+ T cells reduces their ability to be co-stimulated by TLR2 ligands for cytokine production. These include both pro-(IFN-{gamma}, TNF-) and anti-inflammatory cytokines (IL-10). Deletion of TLR2 in T cells did not affect early control but did result in decreased late control of Mtb in the lungs of infected mice. This suggests that T cell co-stimulation by mycobacterial TLR2 ligands in vivo is important for control of infection during the chronic phase of Mtb infection in the lung.

immunology↗

bioRxiv2020

Contribution of noncanonical antigens to virulence and adaptive immunity in human infection with enterotoxigenic E. coli

Enterotoxigenic E. coli (ETEC) contribute significantly to the substantial burden of infectious diarrhea among children living in low and middle income countries. In the absence of a vaccine for ETEC, children succumb to acute dehydration as well as non-diarrheal sequelae related to these infections including malnutrition. The considerable diversity of ETEC genomes has complicated canonical vaccine development approaches focused on a subset of antigens known as colonization factors (CFs). To identify additional conserved immunogens, we mined genomic sequences of 89 ETEC isolates, bioinformatically selected potential surface-exposed pathovar-specific antigens conserved in more than 40% of the genomes (n=118), and assembled the representative proteins onto microarrays, complemented with known or putative colonization factor subunit molecules (n=52), and toxin subunits to interrogate samples from individuals with acute symptomatic ETEC infections. Surprisingly, in this open-aperture approach, we found that immune responses were largely constrained to a small number of antigens including individual colonization factor antigens and EtpA, an extracellular adhesin. In a Bangladeshi cohort of naturally infected children < 2 years of age, both EtpA and a second noncanonical antigen, EatA, elicited significant serologic responses that were associated with protection from symptomatic illness. In addition, children infected with ETEC isolates bearing either etpA or eatA genes were significantly more likely to develop symptomatic disease. These studies support a role for more recently discovered noncanonical antigens in virulence and the development of adaptive immune responses during ETEC infections, findings that may inform vaccine design efforts to complement existing approaches.

microbiology↗

bioRxiv2020

PanGIA: A Metagenomics Analytical Framework for RoutineBiosurveillance and Clinical Pathogen Detection

Metagenomics is emerging as an important tool in biosurveillance, public health, and clinical applications. However, ease-of-use for execution and data analysis remains a barrier-of-entry to the adoption of metagenomics in applied health and forensics settings. In addition, these venues often have more stringent requirements for reporting, accuracy, and precision than the traditional ecological research role of the technology. Here, we present PanGIA (Pan-Genomics for Infectious Agents), a novel bioinformatics analysis platform for hosting, processing, analyzing, and reporting shotgun metagenomics data of complex samples suspected of containing one or more pathogens. PanGIA was developed to address gaps that often preclude clinicians, medical technicians, forensics personnel, or other non-expert end-users from the routine application of metagenomics for pathogen identification. Though primarily designed to detect pathogenic microorganisms within clinical and environmental metagenomics data, PanGIA also serves as an analytical framework for microbial community profiling and comparative metagenomics. To provide statistical confidence in PanGIAs taxonomic assignments, the system provides two independent estimations of probability for species and strain level detection. First, PanGIA integrates coverage data with uniqueness information mapped across each reference genome for a stand-alone determination of confidence for each query sequence at each taxonomy level. Second, if a negative-control sample is provided, PanGIA compares this sample with a corresponding experimental unknown sample and determines a measure of confidence associated with detection above background. An integrated graphical user interface allows interactive interrogation and enables users to summarize multiple sample results by confidence score, normalized read abundance, reference genome linear coverage, depth-of-coverage, RPKM, and other metrics to detect specific organisms-of-interest. Comparison testing of the PanGIA algorithm against a number of recent k-mer, read-mapping, and marker-gene based taxonomy classifiers across various real-world datasets with spiked targets shows superior mean positive predictive value, sensitivity, and specificity. PanGIA can process a five million paired-end read dataset in under 1 hour on commodity computational hardware. The source code and documentation are publicly available at https://github.com/LANL-Bioinformatics/PanGIA or https://github.com/mriglobal/PanGIA. The database for PanGIA can be downloaded from ftp://bioinformatics.mriglobal.org/. The full GUI-based PanGIA analysis environment is available in a Docker container and can be installed from https://hub.docker.com/r/poeli/pangia/.

bioinformatics↗

bioRxiv2020

Critical role for cold shock protein YB-1 in cytokinesis

High levels of the cold shock protein Y-box-binding protein-1, YB-1, are tightly correlated with increased cell proliferation and cancer progression. However, the precise mechanism by which YB-1 regulates proliferation is unknown. Here, we found that YB-1 depletion in several cell lines resulted in cytokinesis failure, multinucleation and an increase in G1 transit time. Rescue experiments indicated that YB-1 was required for completion of cytokinesis. Using confocal imaging of cells undergoing cytokinesis both in vitro and in zebrafish embryos, we found that YB-1 was critical for microtubule organization during cytokinesis. Using mass spectrometry we identified multiple novel phosphorylation sites on YB-1. We show that phosphorylation of YB-1 at multiple serine residues was essential for its function during cytokinesis. Using atomistic modelling we show how multiple phosphorylations alter YB-1 conformation, allowing it to interact with protein partners. Our results establish phosphorylated YB-1 as a critical regulator of cytokinesis, defining for the first time precisely how YB-1 regulates cell division. SUMMARYY-box-binding protein-1, YB-1, is essential for cell division, but it is not clear how it functions. Using live imaging and confocal microscopy we show that YB-1 functions only in the last step of division, specifically being required to initiate cytokinesis.

cell biology↗

bioRxiv2017

Recurrent loss of heterozygosity correlates with clinical outcome in pancreatic neuroendocrine cancer

Pancreatic neuroendocrine tumors (pNETs) are uncommon cancers arising from pancreatic islet cells. Analysis of gene mutation, copy number and RNA expression of 57 sporadic pNETs showed that pNET genomes are dominated by aneuploidy. Remarkably, ~25% of pNETs had genomes characterized by recurrent loss of heterozygosity (LoH) of the same 10 chromosomes, accompanied by bi-allelic MEN1 inactivation, and these cases had generally poor clinical outcome. Another ~25% of all pNETs had chromosome 11 LoH and bi-allelic MEN1 inactivation, lacking the recurrent LoH pattern - these had universally good clinical outcome. Some level of aneuploidy was common, and overall ~80% of pNETs had LoH of [≥]1 chromosome. This aneuploidy led to changes in RNA expression at the level of whole chromosomes and allowed pathogenic germline variants (e.g. ATM) to be expressed unopposed, inactivating downstream tumor suppressor pathways. Some pNETs appear to utilize VHL gene methylation or mutation to activate pseudo-hypoxia. Contrary to expectation neither tumor morphology within well-differentiated pNETs nor single gene mutation had significant associations with clinical outcome, nor did expression of RNAs reflecting the activity of immune, differentiation, proliferative or tumor suppressor pathways. MEN1 was the only statistically significant recurrently mutated driver gene in pNETs. Only one pNET had clearly oncogenic and actionable SNVs (in PTEN and FLCN) confirmed by corroborating RNA expression changes. The two distinct patterns of aneuploidy described here, associated with markedly poor and good clinical outcome respectively, define a novel oncogenic mechanism and the first route to genomic precision oncology for this tumor type.

cancer biology↗