Blair J. Rossetti’s research while affiliated with Emory University and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (12)


Semi-blind sparse affine spectral unmixing of autofluorescence-contaminated micrographs
  • Article

August 2019

·

53 Reads

·

11 Citations

Bioinformatics

Blair J Rossetti

·

Steven A Wilbert

·

·

[...]

·

James G Nagy

Motivation Spectral unmixing methods attempt to determine the concentrations of different fluorophores present at each pixel location in an image by analyzing a set of measured emission spectra. Unmixing algorithms have shown great promise for applications where samples contain many fluorescent labels; however, existing methods perform poorly when confronted with autofluorescence-contaminated images. Results We propose an unmixing algorithm designed to separate fluorophores with overlapping emission spectra from contamination by autofluorescence and background fluorescence. First, we formally define a generalization of the linear mixing model, called the affine mixture model (AMM), that specifically accounts for background fluorescence. Second, we use the AMM to derive an affine nonnegative matrix factorization method for estimating fluorophore endmember spectra from reference images. Lastly, we propose a semi-blind sparse affine spectral unmixing (SSASU) algorithm that uses knowledge of the estimated endmembers to learn the autofluorescence and background fluorescence spectra on a per-image basis. When unmixing real-world spectral images contaminated by autofluorescence, SSASU greatly improved proportion indeterminacy as compared to existing methods for a given relative reconstruction error. Availability and implementation The source code used for this paper was written in Julia and is available with the test data at https://github.com/brossetti/ssasu.


Fig. 1. Comparison of seven Mean and ANMF estimated endmember spectra with fluorometer measurements. The shaded regions represent the fluorometer data, the dotted lines represent the Mean estimates, and the dashed lines represent the ANMF estimates. The gray vertical lines show the wavelength where dichroic mirrors blocked the measurement of emitted light (i.e. locations of missing spectral data).
Fig. 2. Comparison of unmixing performance for SSASU, NLS, and PoissonNMF across ten test images taken from five samples. The relative reconstruction error (top) evaluates each method's ability to reconstruct the observed spectra image. The proportion indeterminacy (bottom) measures the non-orthogonality of the weight matrices and illustrates how well each method separates the fluorophore endmembers in the presence of autofluorescence.
Fig. 3. Montage of unmixed images for NLS (top) and SSASU (bottom). Panels A-P show the unmixed channels for autofluorescence (A, I); S. mitis/DY-415 (B, J); S. salivarius/DY-490 (C, K); Prevotella/ATTO 520 (D, L); Veillonella/ATTO 550 (E, M); Actinomyces/Texas Red-X (F, N); Neisseriaceae/ATTO 620 (G, O); and Rothia/ATTO 655 (H, P). A larger composite view of the non-autofluorescence unmixed channels is shown for NLS in panel Q and for SSASU in panel R. The scale bar in panel R indicates 10 µm.
Fig. 4. Comparison of the autofluorescence endmember estimated from the no-probe control reference image (gray region) to the autofluorescence endmembers learned by SSASU.
Semi-blind sparse affine spectral unmixing of autofluorescence-contaminated micrographs
  • Preprint
  • File available

January 2019

·

125 Reads

Spectral unmixing methods attempt to determine the concentrations of different fluorophores present at each pixel location in an image by analyzing a set of measured emission spectra. Unmixing algorithms have shown great promise for applications where samples contain many fluorescent labels; however, existing methods perform poorly when confronted with autofluorescence-contaminated images. We propose an unmixing algorithm designed to separate fluorophores with overlapping emission spectra from contamination by autofluorescence and background fluorescence. First, we formally define a generalization of the linear mixing model, called the affine mixture model (AMM), that specifically accounts for background fluorescence. Second, we use the AMM to derive an affine nonnegative matrix factorization method for estimating endmember spectra from reference images. Lastly, we propose a semi-blind sparse affine spectral unmixing (SSASU) algorithm that uses knowledge of the estimated endmembers to learn the autofluorescence and background fluorescence spectra on a per-image basis. When unmixing real-world spectral images contaminated by autofluorescence, SSASU was shown to have a similar reconstruction error but greatly improved proportion indeterminacy as compared to existing methods. The source code used for this paper was written in Julia and is available with the test data at https://github.com/brossetti/ssasu.

Download

Clumped Nuclei Segmentation with Adjacent Point Match and Local Shape based Intensity Analysis for Overlapped Nuclei in Fluorescence In-Situ Hybridization Images

August 2018

·

35 Reads

Highly clumped nuclei clusters captured in fluorescence in situ hybridization microscopy images are common histology entities under investigations in a wide spectrum of tissue-related biomedical investigations. Due to their large scale in presence, computer based image analysis is used to facilitate such analysis with improved analysis efficiency and reproducibility. To ensure the quality of downstream biomedical analyses, it is essential to segment clustered nuclei with high quality. However, this presents a technical challenge commonly encountered in a large number of biomedical research, as nuclei are often overlapped due to a high cell density. In this paper, we propose an segmentation algorithm that identifies point pair connection candidates and evaluates adjacent point connections with a formulated ellipse fitting quality indicator. After connection relationships are determined, we recover the resulting dividing paths by following points with specific eigenvalues from Hessian in a constrained searching space. We validate our algorithm with 560 image patches from two classes of tumor regions of seven brain tumor patients. Both qualitative and quantitative experimental results suggest that our algorithm is promising for dividing overlapped nuclei in fluorescence in situ hybridization microscopy images widely used in various biomedical research.


Fig. 1:
Fig. 2:
Clumped Nuclei Segmentation with Adjacent Point Match and Local Shape-Based Intensity Analysis in Fluorescence Microscopy Images

July 2018

·

44 Reads

·

8 Citations

Conference proceedings: ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference

Highly clumped nuclei captured in fluorescence microscopy images are commonly observed in a wide spectrum of tissue-related biomedical investigations. To ensure the quality of downstream biomedical analyses, it is essential to accurately segment clustered nuclei. However, this presents a technical challenge as fluorescence intensity alone is often insufficient for recovering the true nuclei boundaries. In this paper, we propose an segmentation algorithm that identifies point pair connection candidates and evaluates adjacent point connections with a formulated ellipse fitting quality indicator. After connection relationships are determined, we recover the resulting dividing paths by following points with specific eigenvalues from the image Hessian in a constrained searching space. We validate our algorithm with 560 image patches from two classes of tumor regions of seven brain tumor patients. Both qualitative and quantitative experimental results suggest that our algorithm is promising for dividing overlapped nuclei in fluorescence microscopy images widely used in various biomedical research.


GRAPHITE: A Graphical Environment for Scalable in situ Video Tracking of Moving Insects

December 2017

·

25 Reads

·

7 Citations

Methods for measuring animal movement are critical for understanding numerous ecological and evolutionary processes. However, few methods are available for small organisms, and even fewer methods offer consistent individual‐level resolution while remaining affordable, scalable and operable in the field. We describe a low‐cost animal movement tracking method with a user‐friendly graphical interface, called GRAPHITE. Our automated software can quantify motions of insects by offline video analysis of inexpensive and lightweight human‐readable tags attached to individual insects. The integrated graphical editor provides a full‐featured environment for users to review the generated tracking data and make individual‐ or group‐level edits. GRAPHITE is a novel video analysis and graphical editing software ( Matlab v.9.0.0+) that identifies tags in image frames with a minimal false negative rate, links sequences of corresponding tags into “tracks” for each individual insect, infers the tag identifier, and provides a user‐friendly graphical environment for editing tracking data. Users can either batch process raw video data using the full analysis pipeline or execute GRAPHITE modules independently for a tailored analysis. We demonstrate the efficacy of the developed software with a specific application to the movement of honey bees at the entrance of hives. However, this system can be easily modified to track individually marked insects of 3 mm and larger. A notable advantage of this method is its ability to provide easy access to individual‐level tracking data using human‐readable tags.


Preservation of three-dimensional spatial structure in the gut microbiome

November 2017

·

3,789 Reads

·

31 Citations

Preservation of three-dimensional structure in the gut is necessary in order to analyze the spatial organization of the gut microbiota and gut luminal contents. In this study, we evaluated preparation methods for mouse gut with the goal of preserving micron-scale spatial structure while performing fluorescence imaging assays. Our evaluation of embedding methods showed that commonly used media such as Tissue-Tek Optimal Cutting Temperature (OCT) compound, paraffin, and polyester waxes resulted in redistribution of luminal contents. By contrast, a hydrophilic methacrylate resin, Technovit H8100, preserved three-dimensional organization. Our mouse intestinal preparation protocol optimized using the Technovit H8100 embedding method was compatible with microbial fluorescence in situ hybridization (FISH) and other labeling techniques, including immunostaining and staining with both wheat germ agglutinin (WGA) and 4', 6-diamidino-2-phenylindole (DAPI). Mucus could be visualized whether the sample was fixed with paraformaldehyde (PFA) or with Carnoy’s fixative. The protocol optimized in this study enabled simultaneous visualization of micron-scale spatial patterns formed by microbial cells in the mouse intestines along with biogeographical landmarks such as host-derived mucus and food particles.


Preservation of three-dimensional spatial structure in the gut microbiome

August 2017

·

7 Reads

·

4 Citations

Preservation of three-dimensional structure in the gut is necessary in order to analyze the spatial organization of the gut microbiota and gut luminal contents. In this study, we evaluated preparation methods for mouse gut with the goal of preserving micron-scale spatial structure while performing fluorescence imaging assays. Our evaluation of embedding methods showed that commonly used media such as Tissue-Tek Optimal Cutting Temperature (OCT) compound, paraffin, and polyester waxes resulted in redistribution of luminal contents. By contrast, a hydrophilic methacrylate resin, Technovit H8100, preserved three-dimensional organization. Our mouse intestinal preparation protocol optimized using the Technovit H8100 embedding method was compatible with microbial fluorescence in situ hybridization (FISH) and other labeling techniques, including immunostaining and staining with both wheat germ agglutinin (WGA) and 4’,6-diamidino-2-phenylindole (DAPI). Mucus labeling patterns of the samples fixed with paraformaldehyde (PFA) and Carnoy’s fixative were comparable. The protocol optimized in this study enabled simultaneous visualization of micron-scale spatial patterns formed by microbial cells in the mouse intestines along with biogeographical landmarks such as host-derived mucus and food particles.


Dynamic registration for gigapixel serial whole slide images

April 2017

·

54 Reads

·

23 Citations

Proceedings / IEEE International Symposium on Biomedical Imaging: from nano to macro. IEEE International Symposium on Biomedical Imaging

High-throughput serial histology imaging provides a new avenue for the routine study of micro-anatomical structures in a 3D space. However, the emergence of serial whole slide imaging poses a new registration challenge, as the gigapixel image size precludes the direct application of conventional registration techniques. In this paper, we develop a three-stage registration with multi-resolution mapping and propagation method to dynamically produce registered subvolumes from serial whole slide images. We validate our algorithm with gigapixel images of serial brain tumor sections and synthetic image volumes. The qualitative and quantitative assessment results demonstrate the efficacy of our approach and suggest its promise for 3D histology reconstruction analysis.


Biogeography of a human oral microbiome at the micron scale

January 2016

·

364 Reads

·

758 Citations

Proceedings of the National Academy of Sciences

Significance The physiology and ecology of complex microbial communities are strongly dependent on the immediate surroundings of each microbe, including the identity of neighboring microbes; however, information on the micron-scale organization of microbiomes is largely lacking. Using sequencing data combined with spectral fluorescence imaging, we have discovered a multigenus, highly organized microbial consortium in human dental plaque. The spatial structure of the consortium reveals unanticipated interactions and provides a framework for understanding the organization, metabolism, and systems biology of the microbiome and ultimately, its effect on the health of the human host. Our synthesis of high-throughput sequencing data with spatial and structural information shows the informative value of microbial biogeography at the micron scale.


Fig. 1. Detection and quantification of bacterial biofilms on colon tumors. ( A ) FISH of all bacteria (red) on cancers ( Top ), paired normal tissue from patients with CRC ( Middle ), and colonoscopy biopsies from healthy individuals without CRC ( Bottom ). All were counterstained with the nuclear stain, DAPI (blue). The top white brackets demarcate the mucus layer and the bottom white brackets denote the cytoplasm separating the nucleus (blue) of the colorectal epithelium from the mucus layer. PAS stains ( SI Appendix , Fig. S2) further delineate the mucus layer on these samples. ( Insets ) Closeup (100 × ) showing close contact between bacteria and epithelial cells in patient A. The pale, nonpunctate red staining of the mucus layer in patients without biofilms (patient B) represents nonspecific binding to the mucus layer, which is easily demarcated from the bright red punctate staining of the bacteria infiltrating the mucus layer in patients with biofilms. (Scale bars, 50 μ m.) ( B ) Biofilm depth and density measurements from right CRCs/surgical normal pairs ( n = 15), right adenomas/surgical normal pairs ( n = 4), left CRCs/surgical normal pairs ( n = 15), left adenomas/surgical normal pairs ( n = 2), and right/left paired normal colonoscopy biopsies from healthy individuals without CRC ( n = 60). Data displayed as bar and whisker graphs, where line designates the median, boxes the 25/75th percentile, and whiskers the 95th percentile. ( C ) Geographical distribution of tumors (CRC, n = 30, and adenomas, n = 6) with biofilm designation. ( D ) SEM images. ( Left ) Biofilm on a right colon cancer dominated by filamentous bacteria. ( Center ) Biofilm-negative left colon cancer where no bacteria are visualized. ( Right ) Image of bacterial contact with host epithelium (white arrow) on biofilm-covered right colon adenoma. Mixed bacterial morphology (*rods and cocci) is seen. (Scale bars, 2 μ m.) 
Fig. 2. FISH and sequencing analysis of tissue reveal invasive polymicrobial biofilms and tran- sitioning microbial populations. ( A – C ) Multiprobe spectral images of FISH-targeted bacterial groups (40 × ). Bacteroidetes (green), Lachnospiraceae (magenta), Fusobacteria (cyan), Enterobacteriaceae (orange), Bacteroides fragilis (red) are represented within the biofilms, and tissue autofluorescence is white. ( A ) Multigroup bacterial biofilm with invasion of cancer tissue (white arrows). Dotted white line depicts margin between bacterial biofilm and cancer tissue in lower right portion of image. Right cancer with a Fusobacteria dominant polymicrobial biofilm (by sequencing analysis; see text) also containing Bacteroidetes , Lachnospiraceae , and Enterobacteriaceae . Dominant group in left cancer is Bacteroidetes . B. fragilis , Lachnospiraceae , and Fusobacteria are also present. Cancer-invading bacteria represent a subset of biofilm community members. ( B ) Bacterial biofilms on paired surgical normal tissue from CRC patients, comprising Lachnospiraceae , Bacteroidetes , and Enterobacteriaceae . ( C ) Thin bacterial biofilms detected on right ( Bacteroidetes , Lachnospiraceae , and Enterobacteriaceae ) and left ( Bacteroidetes and Lachnospiraceae ) normal colonoscopy biopsies from two different individuals without CRC. ( D , Left ) All bacteria FISH (red) with DAPI nuclear counterstain (blue) of surgically resected normal 
Fig. 3. Biofilms are associated with changes in E-cadherin, IL-6, and Stat3 activation. ( A and B ) Evaluation of E-cadherin and IL-6 by immunofluorescence (green) and activated Stat3 (pStat3, brown nuclei) by immunohistochemistry. Blue, nuclear DAPI counterstain; red, smooth muscle antigen. [Scale bars, 100 μ m (E-cadherin) and 50 μ m (IL-6, pStat3).] ( A ) Normal colonic tissues associated with a biofilm from patients with CRC ( Left ), obtained during surgery, display diminished crypt colonic epithelial cell E-cadherin (white arrows, n = 7 biofilm-positive or -negative tissues) and increased epithelial cell IL-6 (white arrows, n = 13 biofilm-positive or -negative tissues), as well as epithelial cell pStat3 (black arrows, n = 16 biofilm-positive and n = 12 biofilm-negative tissues). Normal colonic tissues without a biofilm from patients with CRC ( Right ), likewise obtained during surgery, display intact E-cadherin. IL-6 and pStat3 are detected in the lamina propria. ( B ) Biofilm-positive colonoscopy biopsies from subjects without CRC ( Left ) display epithelial cell E-cadherin redistribution ( Inset ) and increased tissue IL-6, whereas biofilm-negative colonoscopy biopsies ( Right ) display intact E-cadherin and modest lamina propria IL-6 expression. pStat3 is observed in the lamina propria immune cells in both biofilm-positive and biofilm-negative colonoscopy biopsies. ( C – F ) Quantification of crypt cell E-cadherin (fluorescence intensity), epithelial cell IL-6 [fluorescence intensity and isolated colonic epithelial cells (CEC) by ELISA], and epithelial cell pStat3 (immunohistochemistry) from A are shown in C – F , respectively. Data displayed as bar and whisker graphs, where line designates the median, boxes the 25/75th percentile, and whiskers the 95th percentile ( C and D ) or mean ± SD ( E and F ). Details in SI Appendix , Materials and Methods . SI Appendix , Figs. S10 and S11 show additional E-cadherin, IL-6, and pStat3 quantification from A and B . 
Fig. 4. (A and B) Scoring of Ki67-positive cells from the base of the crypt to the luminal surface. Normal tissues from patients with CRC obtained at surgery (A) with (n = 17) and without (n = 18) a biofilm, as well as normal mucosa from healthy subjects obtained via colonoscopy (B) with (n = 7) and without (n = 10) a biofilm, displayed increased proliferation in a biofilm setting. Data displayed as mean ± SEM in groups based on distance from crypt base (<15 cells, 15-30 cells, >30 cells).
Microbiota organization is a distinct feature of proximal colorectal cancers

December 2014

·

1,295 Reads

·

613 Citations

Proceedings of the National Academy of Sciences

Significance We demonstrate, to our knowledge for the first time, that bacterial biofilms are associated with colorectal cancers, one of the leading malignancies in the United States and abroad. Colon biofilms, dense communities of bacteria encased in a likely complex matrix that contact the colon epithelial cells, are nearly universal on right colon tumors. Most remarkably, biofilm presence correlates with bacterial tissue invasion and changes in tissue biology with enhanced cellular proliferation, a basic feature of oncogenic transformation occurring even in colons without evidence of cancer. Microbiome profiling revealed that biofilm communities on paired normal mucosa cluster with tumor microbiomes but lack distinct taxa differences. This work introduces a previously unidentified concept whereby microbial community structural organization exhibits the potential to contribute to disease progression.


Citations (8)


... These approaches aim to extract the spectral signatures of fluorophores from recorded images and determine the abundance of each fluorophore in every pixel. To tackle unmixing problems in different scenarios, various regularized learning methods have been developed in the literature (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). From a machine learning perspective, these methods essentially represent single-view learning where models are trained and predictions are made based on a single group of features that describes the field of interest, i.e., the emitted spectral profile of the fluorophores. ...

Reference:

A Framework of Multi-View Machine Learning for Biological Spectral Unmixing of Fluorophores with Overlapping Excitation and Emission Spectra
Semi-blind sparse affine spectral unmixing of autofluorescence-contaminated micrographs
  • Citing Article
  • August 2019

Bioinformatics

... The framework surpasses the performance of state-of-the-art detection and segmentation algorithms. Guo et al. (2018) proposed an algorithm for segmentation of overlapped nuclei. The algorithm identifies contenders with point pair connections and evaluates abutted point connections with a contrive ellipse fitting quality criterion. ...

Clumped Nuclei Segmentation with Adjacent Point Match and Local Shape-Based Intensity Analysis in Fluorescence Microscopy Images

Conference proceedings: ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference

... A powerful tool for analyzing collective dynamics of animal behaviors is the utilization of video tracking [14][15][16][17]. This technique is under constant development as it has applicability in a wide range of fields [18]. ...

GRAPHITE: A Graphical Environment for Scalable in situ Video Tracking of Moving Insects
  • Citing Article
  • December 2017

... Notably, PEG chains with MWs of 2 and 5 kDa contributed to the most efficient infiltration of EcN in the mucus at an observation duration of 24 h (Fig. 3a). For detailed visualization of the penetration of EPP, 1 × 10 5 c.f.u.s of bacteria were transferred to a glass-bottom dish pre-spread with a fluorescent mucus layer in which mucin was specifically marked with Alexa Fluor 488-labelled wheat germ agglutinin (AF488-WGA) 31 . LSCM images showed that both uncoated EcN and EP mainly stayed at the top of the mucus layer at 15 min after incubation (Fig. 3b). ...

Preservation of three-dimensional spatial structure in the gut microbiome

... High-resolution WSIs of three serial tissue slides were produced for each patient; one was stained with H&E and two were immunohistochemically stained for Ki67 and pH3. WSI triplets were co-registered at the highest image resolution using our previously developed dynamic co-registration method [18]. For each image triplet, the H&E WSI served as the reference image and the other two IHC images were mapped to the reference image. ...

Dynamic registration for gigapixel serial whole slide images
  • Citing Conference Paper
  • April 2017

Proceedings / IEEE International Symposium on Biomedical Imaging: from nano to macro. IEEE International Symposium on Biomedical Imaging

... We additionally investigated the abundance of several other genera with associations to Streptococcus described in the literature 13 . As other species in addition to Streptococcus may act as early (e.g., Actinomyces, Neisseria, Veillonella, Rothia, Gemella, Granulicatella, Eikenella, Haemophilus, Prevotella) and intermediate (Corynebacterium, Capnocytophaga, Fusobacterium) 14,15 dental plaque colonizers, we investigated whether these taxa have higher relative abundance in samples with low Streptococcus relative abundance. ...

Biogeography of a human oral microbiome at the micron scale
  • Citing Article
  • January 2016

Proceedings of the National Academy of Sciences

... 2025;135(3):e184442 https://doi.org/10.1172/JCI184442 Furthermore, ETBF colonization is common in colon cancer (up to 90%) and epidemiologic studies suggest that it increases the risk of carcinogenesis (75)(76)(77)(78)(79)(80)(81)(82)(83)(84)(85). Following inoculation with ETBF at 5-6 weeks of age, Min mice with intact Hmga1 exhibit poor weight gain and robust distal colon tumorigenesis by 11-12 weeks with a median survival of 17 weeks; by contrast, Min mice with global Hmga1 hemizygosity gain more weight, develop fewer tumors, and exhibit prolonged survival (Figure 4, A-D). ...

Microbiota organization is a distinct feature of proximal colorectal cancers

Proceedings of the National Academy of Sciences

... These communities were distinguished by genetically distinct sequence variants of Neisseria, Veillonella, and Fusobacterium. These bacterial communities were consistent across clustering methods, reproducible in an external cohort [12], and in line with previous work on co-occurrence patterns in oral bacterial communities [26][27][28]. ...

Dynamics of tongue microbial communities with single-nucleotide resolution using oligotyping