Biology subjects

Molnar, J.

Publications and source records attributed to Molnar, J..

7 recordsLinked to original sources

bioRxiv2025

Diploid genome assembly of human fibroblast cell lines enables clone specific variant calling, improved read mapping and accurate phasing

Human cell lines are fundamental tools in biomedical research and are widely used in disease modeling, drug development, and many other domains. Here, we present chromosome-level, phased diploid genome assemblies of two popular human cell lines: the BJ foreskin fibroblast line and the IMR-90 fetal lung fibroblast line. Our high-quality assemblies, generated using long-read and Hi-C sequencing data, reveal substantial structural variation, including more than 50,000 insertions, deletions, duplications, and inversions compared to the recent T2T-CHM13v2.0 reference. Our assemblies provide detailed maps of genetic variation, enabling more accurate variant calling and the ability to phase reads when using newly generated or historical sequencing data on these cell lines or their derivatives. All assemblies and associated data have been made available as a resource for the research community. We envision that diploid genome assembly will become a cornerstone approach for personalized medicine in the near future.

genomics↗

bioRxiv2024

When the pen is mightier than the sword: semi-automatic 2 and 3D image labelling

Data is the driving engine of learning-based algorithms, the creation of which fundamentally determines the performance, accuracy, generalizability and quality of any model or method trained on it. When only skilled or trained personnel can create reliable annotations, assisted software solutions are desirable to reduce the time and effort the expert must spend on labelling. Herein is proposed an automated annotation helper software package in napari that offers multiple methods to assist the annotator in creating object-based labels on 2D or 3D images.

bioinformatics↗

bioRxiv2020

A SNAI2-PEAK1 stromal axis drives progression and lapatinib resistance in HER2-positive breast cancer by supporting a cytokine expression profile that converges on PI3K/Akt signaling

Intercellular mechanisms by which the stromal microenvironment contributes to solid tumor progression and targeted therapy resistance remain poorly understood, presenting significant clinical hurdles. PEAK1 (Pseudopodium-Enriched Atypical Kinase One) is an actin cytoskeleton- and focal adhesion-associated pseudokinase that promotes cell state plasticity and cancer metastasis by mediating growth factor-integrin signaling crosstalk. Here, we determined that stromal PEAK1 expression predicts poor outcomes in HER2-positive breast cancers high in SNAI2 expression and enriched for MSC content. Notably, we identified that mesenchymal stem cells (MSCs) and cancer-associated fibroblasts (CAFs) express high PEAK1 protein levels and MSCs require PEAK1 to potentiate tumorigenesis, lapatinib resistance and metastasis of HER2-positive breast cancer cells. Analysis of PEAK1-dependent secreted factors from MSCs revealed a CCL4-, INHBA- and GDF5-focused network that converged on PI3K/Akt signaling. In this regard, we observed that MSC expression of PEAK1 is required for sustained Akt phosphorylation in neighboring HER2-positive breast cancer cells following lapatinib treatment. Finally, we uncovered a significant correlation between INHBA and PEAK1 expression levels in breast cancer, and that INHBA is an excellent predictor of disease relapse and decreased survival in HER2-positive tumors enriched for PEAK1 expression and MSC content. Importantly, we provide the first evidence that PEAK1 promotes tumorigenic phenotypes through a previously unrecognized SNAI2-PEAK1-INHBA-PI3K/Akt stromal to tumor cell signaling axis. These results establish a new, targetable intercellular mechanism that may be leveraged to improve targeted therapy responses and patient outcomes in breast cancer and other stroma-rich malignancies.

cancer biology↗

bioRxiv2020

Automatic deep learning driven label-free image guided patch clamp system for human and rodent in vitro slice physiology

Patch clamp recording of neurons is a labor-intensive and time-consuming procedure. We have developed a tool that fully automatically performs electrophysiological recordings in label-free tissue slices. The automation covers the detection of cells in label-free images, calibration of the micropipette movement, approach to the cell with the pipette, formation of the whole-cell configuration, and recording. The cell detection is based on deep learning. The model was trained on a new image database of neurons in unlabeled brain tissue slices. The pipette tip detection and approaching phase use image analysis techniques for precise movements. High-quality measurements were performed on hundreds of human and rodent neurons. We also demonstrate that further molecular and anatomical analysis can be performed on the recorded cells. The software has a diary module that automatically logs patch clamp events. Our tool can multiply the number of daily measurements to help brain research. ONE SENTENCE SUMMARYNovel deep learning and image analysis algorithms for automated patch clamp systems to reliably measure neurons in human and rodent brain slices.

neuroscience↗

bioRxiv2019

3D-Cell-Annotator: an open-source active surface tool for single cell segmentation in 3D microscopy images

SummarySegmentation of single cells in microscopy images is one of the major challenges in computational biology. It is the first step of most bioimage analysis tasks, and essential to create training sets for more advanced deep learning approaches. Here, we propose 3D-Cell-Annotator to solve this task using 3D active surfaces together with shape descriptors as prior information in a fully- and semi-automated fashion. The software uses the convenient 3D interface of the widely used Medical Imaging Interaction Toolkit (MITK). Results on 3D biological structures (e.g. spheroids, organoids, embryos) show that the precision of the segmentation reaches the level of a human expert.\n\nAvailability and implementation3D-Cell-Annotator is implemented in CUDA/C++ as a patch for the segmentation module of MITK. The 3D-Cell-Annotator enabled MITK distribution can be downloaded at: www.3D-cell-annotator.org. It works under Windows 64-bit systems and recent Linux distributions even on a consumer level laptop with a CUDA-enabled video card using recent NVIDIA drivers.\n\nContactsfilippo.piccinini@irst.emr.it and horvath.peter@brc.mta.hu

bioinformatics↗

bioRxiv2019

AML ImmunoPET: 64Cu-DOTA-anti-CD33 PET-CT imaging of AML in Xenograft mouse model

PurposeAcute myeloid leukemia (AML) is a highly aggressive form of leukemia that results in poor survival outcomes. Currently, diagnosis and prognosis are based on invasive single-point bone marrow biopsies (iliac crest). There is currently no AML-specific non-invasive imaging method to assess disease in the whole body, including in extramedullary organs, representing an unmet clinical need. About 85-90% of human myeloid leukemia cells express CD33 (an accepted biomarker for AML) cell surface receptors, highlighting CD33 as an ideal candidate for AML immunoPET.\n\nExperimental DesignIn this study, we evaluated if 64Cu-DOTA-anti-CD33 murine monoclonal antibody (mAb) can be used for ImmunoPET-CT imaging of AML in preclinical model. For translational purpose, a humanized anti-CD33 antibody was produced, and after confirming its anti-CD33 PET-CT imaging ability we further explored its therapeutic potential.\n\nResults64Cu-DOTA-anti-CD33 based PET-CT imaging detected CD33+ AML in mice with high sensitivity (95.65%) and specificity (100%). In the entire skeletal system, CD33+ PET activity was significantly high in the bones of AML bearing mice over non-leukemic control mice; femur (p<0.00001), tibia (p=0.0001), humerus (p=0.0014), and lumber spine (p<0.00001). Interestingly, the imaging showed CD33+ PET activity prevelant in epiphysis/diaphysis of the bones, indicating spatial heterogeneity. Anti-CD33 therapy using newly developed humanized anti-CD33 mAb as an ADC (p=0.02) and 225Ac-anti-CD33-RIT (p<0.00001), significantly reduced disease burden over that of respective controls.\n\nConclusionWe have successfully developed a novel anti-CD33 immunoPET-CT based non-invasive imaging tool for diagnosis of AML and its spatial distribution, which indicates a preferential skeletal niche.\n\nTranslational RelevanceAcute myeloid leukemia (AML) exhibits extensive heterogeneity in therapeutic response, even in patients sharing similar characteristics, indicating the inadequacy of currently used biopsy-based prognostic tools. There is a critical need for non-invasive quantitative imaging modality specific to AML, which is able to assess this disease in the whole body, including in extramedullary organs and to longitudinally monitor treatment response. CD33, an accepted AML biomarker is expressed on more than 85-90% of blast cells and Mylotarg (anti-CD33-ADC) is FDA approved immunotherapy for AML. In this preclinical study using anti-human CD33 mAb (64Cu-DOTA-anti-CD33), a CD33 PET-CT imaging modality was developed that detected AML with high sensitivity and specificity. The approach also indicated spatial heterogeneity in leukemic disease, dependent on disease burden, which warrants caution in analyzing results from single point bone marrow biopsies. Therefore, this imaging modality can be used to non-invasively detect AML in humans and also monitor treatment response, an unmet clinical need.

cancer biology↗

bioRxiv2019

A deep learning framework for nucleus segmentation using image style transfer

Single cell segmentation is typically one of the first and most crucial tasks of image-based cellular analysis. We present a deep learning approach aiming towards a truly general method for localizing nuclei across a diverse range of assays and light microscopy modalities. We outperform the 739 methods submitted to the 2018 Data Science Bowl on images representing a variety of realistic conditions, some of which were not represented in the training data. The key to our approach is to adapt our model to unseen and unlabeled data using image style transfer to generate augmented training samples. This allows the model to recognize nuclei in new and different experiments without requiring expert annotations.

bioinformatics↗