analyzing-mitochondrial-morphology-through-simulation-supervised

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

The quantitative analysis of subcellular organelles such as mitochondria in cell fluorescence microscopy images is a demanding task because of the ...

UiT The Arctic University of Norway

This paper presents eigen-analysis of image stack of blinking fluorophores to identify the components that enable super-resolved imaging of blinking fluorophores. Eigen-analysis reveals that the contributions of spatial distribution of fluorophores and their temporal photon emission characteristics can be completely separated. While cross-emitter cross-pixel information of spatial distribution that permits super-resolution is encoded in two matrices, temporal statistics weigh the contribution of these matrices to the measured data. The properties and conditions of exploitation of these matrices are investigated. Con-temporary super-resolution imaging methods that use blinking for super-resolution are studied in the context of the presented analysis. Besides providing insight into the capabilities and limitations of existing super-resolution methods, the analysis shall help in designing better super-resolution techniques that directly exploit these matrices.

Eigen-analysis reveals components supporting super-resolution imaging of blinking fluorophores.

Localization microscopy and multiple signal classification algorithm use temporal stack of image frames of sparse emissions from fluorophores to provide super-resolution images. Localization microscopy localizes emissions in each image independently and later collates the localizations in all the frames, giving same weight to each frame irrespective of its signal-to-noise ratio. This results in a bias towards frames with low signal-to-noise ratio and causes cluttered background in the super-resolved image. User-defined heuristic computational filters are employed to remove a set of localizations in an attempt to overcome this bias. Multiple signal classification performs eigen-decomposition of the entire stack, irrespective of the relative signal-to-noise ratios of the frames, and uses a threshold to classify eigenimages into signal and null subspaces. This results in under-representation of frames with low signal-to-noise ratio in the signal space and over-representation in the null space. Thus, multiple signal classification algorithms is biased against frames with low signal-to-noise ratio resulting into suppression of the corresponding fluorophores. This paper presents techniques to automatically debias localization microscopy and multiple signal classification algorithm of these biases without compromising their resolution and without employing heuristics, user-defined criteria. The effect of debiasing is demonstrated through five datasets of invitro and fixed cell samples.

Non-heuristic automatic techniques for overcoming low signal-to-noise-ratio bias of localization microscopy and multiple signal classification algorithm.

Ultrasound based structural health monitoring of piezoelectric material is challenging if a damage changes at a microscale over time. Classifying geometrically similar damages with a difference in diameter as small as 100 μ m is difficult using conventional sensing and signal analysis approaches. Here, we use an unconventional ultrasound sensing approach that collects information of the entire bulk of the material and investigate the applicability of machine learning approaches for classifying such similar defects. Our results show that appropriate feature design combined with simple k-nearest neighbor classifier can provide up to 98% classification accuracy even though conventional features for time-series data and a variety of classifiers cannot achieve close to 70% accuracy. The newly proposed hybrid feature, which combines frequency domain information in the form of power spectral density and time domain information in the form of sign of slope change, is a suitable feature for achieving the best classification accuracy on this challenging problem.

Classification of Micro-Damage in Piezoelectric Ceramics Using Machine Learning of Ultrasound Signals.

We present an open-source implementation of the fluctuation-based nanoscopy method MUSICAL for ImageJ. This implementation improves the algorithm's computational efficiency and takes advantage of multi-threading to provide orders of magnitude faster reconstructions than the original MATLAB implementation. In addition, the plugin is capable of generating super-resolution videos from large stacks of time-lapse images via an interleaved reconstruction, thus enabling easy-to-use multi-color super-resolution imaging of dynamic systems.

MusiJ: an ImageJ plugin for video nanoscopy.

Image denoising or artefact removal using deep learning is possible in the availability of supervised training dataset acquired in real experiments or synthesized using known noise models. Neither of the conditions can be fulfilled for nanoscopy (super-resolution optical microscopy) images that are generated from microscopy videos through statistical analysis techniques. Due to several physical constraints, a supervised dataset cannot be measured. Further, the non-linear spatio-temporal mixing of data and valuable statistics of fluctuations from fluorescent molecules that compete with noise statistics. Therefore, noise or artefact models in nanoscopy images cannot be explicitly learned. Here, we propose a robust and versatile simulation-supervised training approach of deep learning auto-encoder architectures for the highly challenging nanoscopy images of sub-cellular structures inside biological samples. We show the proof of concept for one nanoscopy method and investigate the scope of generalizability across structures, and nanoscopy algorithms not included during simulation-supervised training. We also investigate a variety of loss functions and learning models and discuss the limitation of existing performance metrics for nanoscopy images. We generate valuable insights for this highly challenging and unsolved problem in nanoscopy, and set the foundation for the application of deep learning problems in nanoscopy for life sciences.

Artefact removal in ground truth deficient fluctuations-based nanoscopy images using deep learning.

Contrast in fluorescence microscopy images allows for the differentiation between different structures by their difference in intensities. However, factors such as point-spread function and noise may reduce it, affecting its interpretability. We identified that fluctuation of emitters in a stack of images can be exploited to achieve increased contrast when compared to the average and Richardson-Lucy deconvolution. We tested our methods on four increasingly challenging samples including tissue, in which case results were comparable to the ones obtained by structured illumination microscopy in terms of contrast.

Deriving high contrast fluorescence microscopy images through low contrast noisy image stacks.

Mitochondria are essential energy-providing organelles of particular importance in energy-demanding tissue such as the heart. The production of mitochondria-derived vesicles (MDVs) is a cellular mechanism by which cells ensure a healthy pool of mitochondria. These vesicles are small and fast-moving objects not easily captured by imaging. In this work, we have tested the ability of the optical super-resolution technique 3DSIM to capture high-resolution images of MDVs. We optimized the imaging conditions both for high-speed video microscopy and fixed-cell imaging and analysis. From the 3DSIM videos, we observed an abundance of MDVs and many dynamic mitochondrial tubules. The density of MDVs in cells was compared for cells under normal growth conditions and cells during metabolic perturbation. Our results indicate a higher abundance of MDVs in H9c2 cells during glucose deprivation compared with cells under normal growth conditions. Furthermore, the results reveal a large untapped potential of 3DSIM in MDV research.

Mitochondrial dynamics and quantification of mitochondria-derived vesicles in cardiomyoblasts using structured illumination microscopy.

Visualization of three-dimensional (3D) morphological changes in the subcellular structures of a biological specimen is a major challenge in life science. Here, we present an integrated chip-based optical nanoscopy combined with quantitative phase microscopy (QPM) to obtain 3D morphology of liver sinusoidal endothelial cells (LSEC). LSEC have unique morphology with small nanopores (50-300 nm in diameter) in the plasma membrane, called fenestrations. The fenestrations are grouped in discrete clusters, which are around 100 to 200 nm thick. Thus, imaging and quantification of fenestrations and sieve plate thickness require resolution and sensitivity of sub-100 nm along both the lateral and the axial directions, respectively. In chip-based nanoscopy, the optical waveguides are used both for hosting and illuminating the sample. The fluorescence signal is captured by an upright microscope, which is converted into a Linnik-type interferometer to sequentially acquire both superresolved images and phase information of the sample. The multimodal microscope provided an estimate of the fenestration diameter of 119 ± 53 nm and average thickness of the sieve plates of 136.6 ± 42.4 nm, assuming the constant refractive index of cell membrane to be 1.38. Further, LSEC were treated with cytochalasin B to demonstrate the possibility of precise detection in the cell height. The mean phase value of the fenestrated area in normal and treated cells was found to be 161 ± 50 mrad and 109 ± 49 mrad, respectively. The proposed multimodal technique offers nanoscale visualization of both the lateral size and the thickness map, which would be of broader interest in the fields of cell biology and bioimaging.

Multimodal on-chip nanoscopy and quantitative phase imaging reveals the nanoscale morphology of liver sinusoidal endothelial cells.

Histology involves the observation of structural features in tissues using a microscope. While diffraction-limited optical microscopes are commonly used in histological investigations, their resolving capabilities are insufficient to visualize details at subcellular level. Although a novel set of super-resolution optical microscopy techniques can fulfill the resolution demands in such cases, the system complexity, high operating cost, lack of multi-modality, and low-throughput imaging of these methods limit their wide adoption for histological analysis. In this study, we introduce the photonic chip as a feasible high-throughput microscopy platform for super-resolution imaging of histological samples. Using cryopreserved ultrathin tissue sections of human placenta, mouse kidney, pig heart, and zebrafish eye retina prepared by the Tokuyasu method, we demonstrate diverse imaging capabilities of the photonic chip including total internal reflection fluorescence microscopy, intensity fluctuation-based optical nanoscopy, single-molecule localization microscopy, and correlative light-electron microscopy. Our results validate the photonic chip as a feasible imaging platform for tissue sections and pave the way for the adoption of super-resolution high-throughput multimodal analysis of cryopreserved tissue samples both in research and clinical settings.

Chip-based multimodal super-resolution microscopy for histological investigations of cryopreserved tissue sections.

This three-dimensional structured illumination microscopy (3DSIM) dataset was generated to highlight the suitability of 3DSIM to investigate mitochondria-derived vesicles (MDVs) in H9c2 cardiomyoblasts in living or fixed cells. MDVs act as a mitochondria quality control mechanism. The cells were stably expressing the tandem-tag eGFP-mCherry-OMP25-TM (outer mitochondrial membrane) which can be used as a sensor for acidity. A part of the dataset is showing correlative imaging of lysosomes labeled using LysoTracker in fixed and living cells. The cells were cultivated in either normal or glucose-deprived medium containing galactose. The resulting 3DSIM data were of high quality and can be used to undertake a variety of studies. Interestingly, many dynamic tubules derived from mitochondria are visible in the 3DSIM videos under both glucose and galactose-adapted growth conditions. As the raw 3DSIM data, optical parameters, and reconstructed 3DSIM images are provided, the data is especially suitable for use in the development of SIM reconstruction algorithms, bioimage analysis methods, and for biological studies of mitochondria.

Three-dimensional structured illumination microscopy data of mitochondria and lysosomes in cardiomyoblasts under normal and galactose-adapted conditions.

Mitochondria play a crucial role in cellular metabolism. This paper presents a novel method to visualize mitochondria in living cells without the use of fluorescent markers. We propose a physics-guided deep learning approach for obtaining virtually labeled micrographs of mitochondria from bright-field images. We integrate a microscope's point spread function in the learning of an adversarial neural network for improving virtual labeling. We show results (average Pearson correlation 0.86) significantly better than what was achieved by state-of-the-art (0.71) for virtual labeling of mitochondria. We also provide new insights into the virtual labeling problem and suggest additional metrics for quality assessment. The results show that our virtual labeling approach is a powerful way of segmenting and tracking individual mitochondria in bright-field images, results previously achievable only for fluorescently labeled mitochondria.

Krishna Agarwal

Department of Physics and Technology

UiT The Arctic University of Norway

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Krishna Agarwal

.css-o250i3{font-size:18px;font-family:Open Sans,sans-serif;line-height:21.6px;}Department of Physics and Technology

UiT The Arctic University of Norway

Analyzing Mitochondrial Morphology Through Simulation Supervised Learning

Department of Physics and Technology