OBJECTIVES: This article focuses on the detection of cells in low-contrast brightfield microscopy images; in our case, it is chronic lymphocytic leukaemia cells. The automatic detection of cells from brightfield time-lapse microscopic images brings new opportunities in cell morphology and migration studies; to achieve the desired results, it is advisable to use state-of-the-art image segmentation methods that not only detect the cell but also detect its boundaries with the highest possible accuracy, thus defining its shape and dimensions. METHODS: We compared eight state-of-the-art neural network architectures with different backbone encoders for image data segmentation, namely U-net, U-net++, the Pyramid Attention Network, the Multi-Attention Network, LinkNet, the Feature Pyramid Network, DeepLabV3, and DeepLabV3+. The training process involved training each of these networks for 1000 epochs using the PyTorch and PyTorch Lightning libraries. For instance segmentation, the watershed algorithm and three-class image semantic segmentation were used. We also used StarDist, a deep learning-based tool for object detection with star-convex shapes. RESULTS: The optimal combination for semantic segmentation was the U-net++ architecture with a ResNeSt-269 background with a data set intersection over a union score of 0.8902. For the cell characteristics examined (area, circularity, solidity, perimeter, radius, and shape index), the difference in mean value using different chronic lymphocytic leukaemia cell segmentation approaches appeared to be statistically significant (Mann-Whitney U test, P < .0001). CONCLUSION: We found that overall, the algorithms demonstrate equal agreement with ground truth, but with the comparison, it can be seen that the different approaches prefer different morphological features of the cells. Consequently, choosing the most suitable method for instance-based cell segmentation depends on the particular application, namely, the specific cellular traits being investigated.
- Klíčová slova
- Cell detection, U-net++, cell segmentation, chronic lymphocytic leukaemia cells, image analysis,
- Publikační typ
- časopisecké články MeSH
MOTIVATION: Optical mapping is a complementary technology to traditional DNA sequencing technologies, such as next-generation sequencing (NGS). It provides genome-wide, high-resolution restriction maps from single, stained molecules of DNA. It can be used to detect large and small structural variants, copy number variations and complex rearrangements. Optical mapping is affected by different kinds of errors in comparison with traditional DNA sequencing technologies. It is important to understand the source of these errors and how they affect the obtained data. This article proposes a novel approach to modeling errors in the data obtained from the Bionano Genomics Inc. Saphyr system with Direct Label and Stain (DLS) chemistry. Some studies have already addressed this issue for older instruments with nicking enzymes, but we are unaware of a study that addresses this new system. RESULTS: The main result is a framework for studying errors in the data obtained from the Saphyr instrument with DLS chemistry. The framework's main component is a simulation that computes how major sources of errors for this instrument (a false site, a missing site and resolution errors) affect the distribution of fragment lengths in optical maps. The simulation is parametrized by variables describing these errors and we are using a differential evolution algorithm to evaluate parameters that best fit the data from the instrument. Results of the experiments manifest that this approach can be used to study errors in the optical mapping data analysis. AVAILABILITY AND IMPLEMENTATION: Source codes supporting the presented results are available at: https://github.com/mvasinek/olgen-om-error-prediction. The data underlying this article are available on the Bionano Genomics Inc. website, at: https://bionanogenomics.com/library/datasets/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.
- Publikační typ
- časopisecké články MeSH
Extramedullary disease (EMM) represents a rare, aggressive and mostly resistant phenotype of multiple myeloma (MM). EMM is frequently associated with high-risk cytogenetics, but their complex genomic architecture is largely unexplored. We used whole-genome optical mapping (Saphyr, Bionano Genomics) to analyse the genomic architecture of CD138+ cells isolated from bone-marrow aspirates from an unselected cohort of newly diagnosed patients with EMM (n = 4) and intramedullary MM (n = 7). Large intrachromosomal rearrangements (> 5 Mbp) within chromosome 1 were detected in all EMM samples. These rearrangements, predominantly deletions with/without inversions, encompassed hundreds of genes and led to changes in the gene copy number on large regions of chromosome 1. Compared with intramedullary MM, EMM was characterised by more deletions (size range of 500 bp-50 kbp) and fewer interchromosomal translocations, and two EMM samples had copy number loss in the 17p13 region. Widespread genomic heterogeneity and novel aberrations in the high-risk IGH/IGK/IGL, 8q24 and 13q14 regions were detected in individual patients but were not specific to EMM/MM. Our pilot study revealed an association of chromosome 1 abnormalities in bone marrow myeloma cells with extramedullary progression. Optical mapping showed the potential for refining the complex genomic architecture in MM and its phenotypes.
- MeSH
- buňky kostní dřeně patologie MeSH
- celogenomová asociační studie metody MeSH
- chromozomální aberace * MeSH
- cytogenetické vyšetření metody MeSH
- kohortové studie MeSH
- kostní dřeň diagnostické zobrazování metabolismus patologie MeSH
- lidé středního věku MeSH
- lidé MeSH
- lidské chromozomy, pár 1 * genetika MeSH
- mnohočetný myelom genetika patologie MeSH
- pilotní projekty MeSH
- senioři nad 80 let MeSH
- senioři MeSH
- Check Tag
- lidé středního věku MeSH
- lidé MeSH
- mužské pohlaví MeSH
- senioři nad 80 let MeSH
- senioři MeSH
- ženské pohlaví MeSH
- Publikační typ
- časopisecké články MeSH
- práce podpořená grantem MeSH
- Geografické názvy
- Česká republika MeSH
The insufficient standardization of diagnostic next-generation sequencing (NGS) still limits its implementation in clinical practice, with the correct detection of mutations at low variant allele frequencies (VAF) facing particular challenges. We address here the standardization of sequencing coverage depth in order to minimize the probability of false positive and false negative results, the latter being underestimated in clinical NGS. There is currently no consensus on the minimum coverage depth, and so each laboratory has to set its own parameters. To assist laboratories with the determination of the minimum coverage parameters, we provide here a user-friendly coverage calculator. Using the sequencing error only, we recommend a minimum depth of coverage of 1,650 together with a threshold of at least 30 mutated reads for a targeted NGS mutation analysis of ≥3% VAF, based on the binomial probability distribution. Moreover, our calculator also allows adding assay-specific errors occurring during DNA processing and library preparation, thus calculating with an overall error of a specific NGS assay. The estimation of correct coverage depth is recommended as a starting point when assessing thresholds of NGS assay. Our study also points to the need for guidance regarding the minimum technical requirements, which based on our experience should include the limit of detection (LOD), overall NGS assay error, input, source and quality of DNA, coverage depth, number of variant supporting reads, and total number of target reads covering variant region. Further studies are needed to define the minimum technical requirements and its reporting in diagnostic NGS.
- Klíčová slova
- TP53 gene, coverage depth calculator, next-generation sequencing, sequencing error, small subclones, variant allele frequency (VAF),
- Publikační typ
- časopisecké články MeSH
