Dataset Information

Robust blind spectral unmixing for fluorescence microscopy using unsupervised learning.

ABSTRACT: Due to the overlapping emission spectra of fluorophores, fluorescence microscopy images often have bleed-through problems, leading to a false positive detection. This problem is almost unavoidable when the samples are labeled with three or more fluorophores, and the situation is complicated even further when imaged under a multiphoton microscope. Several methods have been developed and commonly used by biologists for fluorescence microscopy spectral unmixing, such as linear unmixing, non-negative matrix factorization, deconvolution, and principal component analysis. However, they either require pre-knowledge of emission spectra or restrict the number of fluorophores to be the same as detection channels, which highly limits the real-world applications of those spectral unmixing methods. In this paper, we developed a robust and flexible spectral unmixing method: Learning Unsupervised Means of Spectra (LUMoS), which uses an unsupervised machine learning clustering method to learn individual fluorophores' spectral signatures from mixed images, and blindly separate channels without restrictions on the number of fluorophores that can be imaged. This method highly expands the hardware capability of two-photon microscopy to simultaneously image more fluorophores than is possible with instrumentation alone. Experimental and simulated results demonstrated the robustness of LUMoS in multi-channel separations of two-photon microscopy images. We also extended the application of this method to background/autofluorescence removal and colocalization analysis. Lastly, we integrated this tool into ImageJ to offer an easy to use spectral unmixing tool for fluorescence imaging. LUMoS allows us to gain a higher spectral resolution and obtain a cleaner image without the need to upgrade the imaging hardware capabilities.

SUBMITTER: McRae TD

PROVIDER: S-EPMC6886781 | biostudies-literature | 2019

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Robust blind spectral unmixing for fluorescence microscopy using unsupervised learning.

McRae Tristan D TD Oleksyn David D Miller Jim J Gao Yu-Rong YR

PloS one 20191202 12

Due to the overlapping emission spectra of fluorophores, fluorescence microscopy images often have bleed-through problems, leading to a false positive detection. This problem is almost unavoidable when the samples are labeled with three or more fluorophores, and the situation is complicated even further when imaged under a multiphoton microscope. Several methods have been developed and commonly used by biologists for fluorescence microscopy spectral unmixing, such as linear unmixing, non-negativ ...[more]

PMID: 31790435

Similar Datasets

Project description:BackgroundLiving tissues contain a range of intrinsic fluorophores and sources of second harmonic generation which provide contrast that can be exploited for fresh tissue imaging. Microscopic imaging of fresh tissue samples can circumvent the cost and time associated with conventional histology. Further, intrinsic contrast can provide rich information about a tissue's composition, structure and function, and opens the potential for in-vivo imaging without the need for contrast agents.Methodology/principal findingsIn this study, we used hyperspectral two-photon microscopy to explore the characteristics of both normal and diseased gastrointestinal (GI) tissues, relying only on their endogenous fluorescence and second harmonic generation to provide contrast. We obtained hyperspectral data at subcellular resolution by acquiring images over a range of two-photon excitation wavelengths, and found excitation spectral signatures of specific tissue types based on our ability to clearly visualize morphology. We present the two-photon excitation spectral properties of four major tissue types that are present throughout the GI tract: epithelium, lamina propria, collagen, and lymphatic tissue. Using these four excitation signatures as basis spectra, linear unmixing strategies were applied to hyperspectral data sets of both normal and neoplastic tissue acquired in the colon and small intestine. Our results show that hyperspectral unmixing with excitation spectra allows segmentation, showing promise for blind identification of tissue types within a field of view, analogous to specific staining in conventional histology. The intrinsic spectral signatures of these tissue types provide information relating to their biochemical composition.Conclusions/significanceThese results suggest hyperspectral two-photon microscopy could provide an alternative to conventional histology either for in-situ imaging, or intraoperative 'instant histology' of fresh tissue biopsies.

Project description:BackgroundThis paper introduces a new constrained model and the corresponding algorithm, called unsupervised Bayesian linear unmixing (uBLU), to identify biological signatures from high dimensional assays like gene expression microarrays. The basis for uBLU is a Bayesian model for the data samples which are represented as an additive mixture of random positive gene signatures, called factors, with random positive mixing coefficients, called factor scores, that specify the relative contribution of each signature to a specific sample. The particularity of the proposed method is that uBLU constrains the factor loadings to be non-negative and the factor scores to be probability distributions over the factors. Furthermore, it also provides estimates of the number of factors. A Gibbs sampling strategy is adopted here to generate random samples according to the posterior distribution of the factors, factor scores, and number of factors. These samples are then used to estimate all the unknown parameters.ResultsFirstly, the proposed uBLU method is applied to several simulated datasets with known ground truth and compared with previous factor decomposition methods, such as principal component analysis (PCA), non negative matrix factorization (NMF), Bayesian factor regression modeling (BFRM), and the gradient-based algorithm for general matrix factorization (GB-GMF). Secondly, we illustrate the application of uBLU on a real time-evolving gene expression dataset from a recent viral challenge study in which individuals have been inoculated with influenza A/H3N2/Wisconsin. We show that the uBLU method significantly outperforms the other methods on the simulated and real data sets considered here.ConclusionsThe results obtained on synthetic and real data illustrate the accuracy of the proposed uBLU method when compared to other factor decomposition methods from the literature (PCA, NMF, BFRM, and GB-GMF). The uBLU method identifies an inflammatory component closely associated with clinical symptom scores collected during the study. Using a constrained model allows recovery of all the inflammatory genes in a single factor.

Project description:PurposeImage quality of positron emission tomography (PET) is limited by various physical degradation factors. Our study aims to perform PET image denoising by utilizing prior information from the same patient. The proposed method is based on unsupervised deep learning, where no training pairs are needed.MethodsIn this method, the prior high-quality image from the patient was employed as the network input and the noisy PET image itself was treated as the training label. Constrained by the network structure and the prior image input, the network was trained to learn the intrinsic structure information from the noisy image and output a restored PET image. To validate the performance of the proposed method, a computer simulation study based on the BrainWeb phantom was first performed. A 68Ga-PRGD2 PET/CT dataset containing 10 patients and a 18F-FDG PET/MR dataset containing 30 patients were later on used for clinical data evaluation. The Gaussian, non-local mean (NLM) using CT/MR image as priors, BM4D, and Deep Decoder methods were included as reference methods. The contrast-to-noise ratio (CNR) improvements were used to rank different methods based on Wilcoxon signed-rank test.ResultsFor the simulation study, contrast recovery coefficient (CRC) vs. standard deviation (STD) curves showed that the proposed method achieved the best performance regarding the bias-variance tradeoff. For the clinical PET/CT dataset, the proposed method achieved the highest CNR improvement ratio (53.35% ± 21.78%), compared with the Gaussian (12.64% ± 6.15%, P = 0.002), NLM guided by CT (24.35% ± 16.30%, P = 0.002), BM4D (38.31% ± 20.26%, P = 0.002), and Deep Decoder (41.67% ± 22.28%, P = 0.002) methods. For the clinical PET/MR dataset, the CNR improvement ratio of the proposed method achieved 46.80% ± 25.23%, higher than the Gaussian (18.16% ± 10.02%, P < 0.0001), NLM guided by MR (25.36% ± 19.48%, P < 0.0001), BM4D (37.02% ± 21.38%, P < 0.0001), and Deep Decoder (30.03% ± 20.64%, P < 0.0001) methods. Restored images for all the datasets demonstrate that the proposed method can effectively smooth out the noise while recovering image details.ConclusionThe proposed unsupervised deep learning framework provides excellent image restoration effects, outperforming the Gaussian, NLM methods, BM4D, and Deep Decoder methods.

Dataset Information

Robust blind spectral unmixing for fluorescence microscopy using unsupervised learning.

Publications

Robust blind spectral unmixing for fluorescence microscopy using unsupervised learning.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure