Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

GLAD: GLycan Array Dashboard, a visual analytics tool for glycan microarrays

GLAD: GLycan Array Dashboard, a visual analytics tool for glycan microarrays Motivation: Traditional glycan microarray data is typically presented as excel files with limited visualization and interactivity. Thus, comparisons and analysis of glycan array data have been diffi- cult, and there is need for a tool to facilitate data mining of glycan array data. Results: GLAD (GLycan Array Dashboard) is a web-based tool to visualize, analyze, present and mine glycan microarray data. GLAD allows users to input multiple data files to create comparisons. GLAD extends the capability of the microarray data to produce more comparative visualizations in the form of grouped bar charts, heatmaps, calendar heatmaps, force graphs and correlation maps in order to analyze broad sets of samples. Additionally, it allows users to filter, sort and normalize the data and view glycan structures in an interactive manner, to facilitate faster visual data mining. Availability and implementation: GLAD is freely available for use on the Web at https://glycotool kit.com/Tools/GLAD/ with all major modern browsers (Edge, Firefox, Chrome, Safari). Contact: aymehta@bidmc.harvard.edu or rcummin1@bidmc.harvard.edu Supplementary information: Full documentation and video tutorials for GLAD can be found on https://glycotoolkit.com/GLAD. 1 Introduction Out of the four major macromolecules within human cells, glycans Functional Glycomics (NCFG) contain synthetic glycans of known (oli-gosaccharides) are by far the most complicated as they can be structure and are printed on functionalized glass microarray surfaces made up of multiple monosaccharide units, having 3–5 possible sites (Heimburg-Molinaro et al., 2011). Such arrays can be utilized to of attachments, with 2 anomeric configurations (a or b) for glycosid- screen Glycan Binding Proteins (GBPs) and biological samples (e.g. ic bonds and range in size from one- to >20-mers long (Seeberger, serum) using very small sample volumes to define their binding spe- 2015). Thus, it is not just the diversity in the molecules and length cificities (Liang and Wu, 2009). To date, data for >3900 experi- of the glycan, but also the variety of linkages and possibility of ments has been made public by the CFG, with more data being branching, which increases the molecular complexity of glycans added (http://www.functionalglycomics.org/). orders of magnitudes higher than other macromolecules. Glycans Data visualization is an important step in exploratory data ana- lysis, and should be made comprehensive but less tedious (Shelly, play crucial roles in a number of biological functions and patholo- 1996). Tools to visualize glycan array data and binding patterns are gies such as immunity, host-pathogen interactions, adhesion and cell signaling via binding to glycan binding proteins (Cummings and limited. There have been a few glycan array database tools created Pierce, 2014; Varki, 2017; Varki et al., 2015). Studying interactions previously, for example, CFG database (Venkataraman et al., that involve glycans can potentially lead to a better understanding of 2015), GlycoPattern (Agravat et al., 2014), GlycanBinder and biology for potential diagnostics and therapeutics. GlycoSearch (Kletter et al., 2015; Porter et al., 2010). These resour- Defined glycan arrays such as those produced by the Consortium ces provide large amounts of raw data or automated presentation of for Functional Glycomics (CFG) and the National Center for binding motifs. However, the software have limited data V The Author(s) 2019. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 3536 Downloaded from https://academic.oup.com/bioinformatics/article-abstract/35/18/3536/5305635 by Ed 'DeepDyve' Gillespie user on 18 September 2019 GLAD: GLycan Array Dashboard 3537 visualization features (bar graphs and heatmaps) and are restrictive found in the GLAD documentation online along with video snippets with respect to usage (need login or special access), thus limiting and tutorials to guide the user on how to use the tools. In addition, we their use in visual analytics. To facilitate the data visualization and have put up an example demonstrating the ability to use GLAD to dif- visual comparison of glycan array data, we present here a new tool, ferentiate between Galectin-3, Galectin-3C and Galectin-7 online. GLAD (GLycan Array Dashboard), which is a browser-based tool to visualize and mine glycan array data. GLAD was developed using 3 Results and discussion a combination of JavaScript (JS) libraries including D3.js, jQuery, Lodash, Bootstrap and Select2js. As GLAD is mainly coded in JS, it GLAD is a unique tool offered to glycoscientists to assist in mining removes the need for users to upload their data to any server, and and comparing glycan array data in an easy and private manner. hence keeps user data and analysis private on their local site. This GLAD can be extremely useful in uncovering hidden relationships significantly reduces the need to maintain a high-performance ser- between array datasets, which was previously unexplored. The abil- ver. Furthermore, the data visualization includes not only bar graphs ity to visualize glycan array data, alongside glycan structure and the and heatmaps, but also calendar heatmaps, force graphs and correl- ability to filter the data based on user-defined criteria enables the ation maps. All visualizations are interactive and in SVG format for user to mine data in an efficient and thorough manner. In addition, the best user experience. GLAD also allows users to filter, sort and the selection files which are saved in the tool can be attached to normalize data to accentuate key data and binding relationships. manuscripts as Supplementary Information for reviewers and read- Overall, GLAD helps to make glycan array data analysis simpler ers to have a more interactive view of the data. The tool is constant- and more easily accessible to a larger community. ly being updated to add new features as requested by internal users of the NCFG, which provides glycan microarray assays as a service to the scientific community, and is open to suggestions from the gly- 2 Materials and methods coscience community. Additionally, newer data visualizations are constantly being investigated. In the future, we intend to host a data- The general workflow for GLAD is as follows (Fig. 1): 1. Input data base, which could feed data that is public into the GLAD interface as tab-delimited text files in the correct format. 2. Visualize and se- for visualization. This will allow users to analyze and compare their lect data for display using bar charts (Stage-1), or, alternatively add data along with publicly available data. the entire data file to the selections. 3. Save/Load the selections from multiple experiments. This allows the user to perform their analysis locally and resume at any time. Filter, Sort or Normalize the selec- Acknowledgements tions options allow the user to mine the data to analyze binding pat- terns. 4. Visualize the data selected from various experiments in one The authors would like to thank all members of the National Center for Functional Glycomics for their valuable feedback, especially Chao Gao, of the Stage-2 visualizations, i.e.: Lauren Byrd-Leotis, Tanya McKitrick, Jamie Heimburg-Molinaro and Alyssa Grouped Bar Charts- useful to compare small datasets. McQuillan. Heatmaps- useful to compare data for large numbers of samples across small numbers of glycans. Calendar Heatmaps- useful to compare data for large numbers Funding of glycans across multiple experiments. This work was supported by National Institutes of Health Grants Force Graphs- useful to visualize any network-like binding pat- P41GM103694 and U01GM125267. terns and identify common binding glycans among samples. Conflict of Interest: none declared. Correlation plots- useful to compare and quantify similarities or differences in binding patterns for a set of experiments, based on the generated Pearson Coefficient (r) values. References Bubble Box plots- allows users to get statistical values as box Agravat,S.B. et al. (2014) GlycoPattern: a web platform for glycan array min- plots and mean values with scatter-plot like data. ing. Bioinformatics, 30, 3417–3418. Cummings,R.D. and Pierce,J.M. (2014) The challenge and promise of glyco- The users can save and export the figures as SVG files. Users can mics. Chem. Biol., 21, 1–15. view the glycan structure in the Glycan Structure drawer at any time Heimburg-Molinaro,J. et al. (2011) Preparation and analysis of glycan micro- by hovering the mouse over data points. Full documentation can be arrays. Curr. Protoc. Protein Sci., 10, 12.10.1–12.10.29. Kletter,D. et al. (eds.) (2015) Glycoscience: Biology and Medicine. Springer, Tokyo, Japan, pp. 61–68. Liang,C.H. and Wu,C.Y. (2009) Glycan array: a powerful tool for glycomics studies. Exp. Rev. Proteomics, 6, 631–645. Porter,A. et al. (2010) A motif-based analysis of glycan array data to determine the specificities of glycan-binding proteins. Glycobiology, 20, 369–380. Seeberger,P.H. (2015) Monosaccharide diversity. In: Varki,A. et al. (eds.) Essentials of Glycobiology. Cold Spring Harbor, New York, pp. 19–30. Shelly,M.A. (1996) Exploratory data analysis: data visualization or torture? Infect. Control Hosp. Epidemiol., 17, 605–612. Varki,A. (2017) Biological roles of glycans. Glycobiology, 27, 3–49. Varki,A. et al. (eds.) (2015) Essentials of Glycobiology. Cold Spring Harbor, New York. pp. 77–88. Fig. 1. GLAD workflow showing how the user can input, manipulate and visu- Venkataraman,M. et al. (2015) Glycan array data management at consortium alize glycan array data for functional glycomics. Methods Mol. Biol., 1273, 181–190. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Bioinformatics Oxford University Press

GLAD: GLycan Array Dashboard, a visual analytics tool for glycan microarrays

Bioinformatics , Volume 35 (18): 2 – Feb 1, 2019

Loading next page...
 
/lp/oxford-university-press/glad-glycan-array-dashboard-a-visual-analytics-tool-for-glycan-jLd0AsDkjV

References (13)

Publisher
Oxford University Press
Copyright
© The Author(s) 2019. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1367-4803
eISSN
1460-2059
DOI
10.1093/bioinformatics/btz075
Publisher site
See Article on Publisher Site

Abstract

Motivation: Traditional glycan microarray data is typically presented as excel files with limited visualization and interactivity. Thus, comparisons and analysis of glycan array data have been diffi- cult, and there is need for a tool to facilitate data mining of glycan array data. Results: GLAD (GLycan Array Dashboard) is a web-based tool to visualize, analyze, present and mine glycan microarray data. GLAD allows users to input multiple data files to create comparisons. GLAD extends the capability of the microarray data to produce more comparative visualizations in the form of grouped bar charts, heatmaps, calendar heatmaps, force graphs and correlation maps in order to analyze broad sets of samples. Additionally, it allows users to filter, sort and normalize the data and view glycan structures in an interactive manner, to facilitate faster visual data mining. Availability and implementation: GLAD is freely available for use on the Web at https://glycotool kit.com/Tools/GLAD/ with all major modern browsers (Edge, Firefox, Chrome, Safari). Contact: aymehta@bidmc.harvard.edu or rcummin1@bidmc.harvard.edu Supplementary information: Full documentation and video tutorials for GLAD can be found on https://glycotoolkit.com/GLAD. 1 Introduction Out of the four major macromolecules within human cells, glycans Functional Glycomics (NCFG) contain synthetic glycans of known (oli-gosaccharides) are by far the most complicated as they can be structure and are printed on functionalized glass microarray surfaces made up of multiple monosaccharide units, having 3–5 possible sites (Heimburg-Molinaro et al., 2011). Such arrays can be utilized to of attachments, with 2 anomeric configurations (a or b) for glycosid- screen Glycan Binding Proteins (GBPs) and biological samples (e.g. ic bonds and range in size from one- to >20-mers long (Seeberger, serum) using very small sample volumes to define their binding spe- 2015). Thus, it is not just the diversity in the molecules and length cificities (Liang and Wu, 2009). To date, data for >3900 experi- of the glycan, but also the variety of linkages and possibility of ments has been made public by the CFG, with more data being branching, which increases the molecular complexity of glycans added (http://www.functionalglycomics.org/). orders of magnitudes higher than other macromolecules. Glycans Data visualization is an important step in exploratory data ana- lysis, and should be made comprehensive but less tedious (Shelly, play crucial roles in a number of biological functions and patholo- 1996). Tools to visualize glycan array data and binding patterns are gies such as immunity, host-pathogen interactions, adhesion and cell signaling via binding to glycan binding proteins (Cummings and limited. There have been a few glycan array database tools created Pierce, 2014; Varki, 2017; Varki et al., 2015). Studying interactions previously, for example, CFG database (Venkataraman et al., that involve glycans can potentially lead to a better understanding of 2015), GlycoPattern (Agravat et al., 2014), GlycanBinder and biology for potential diagnostics and therapeutics. GlycoSearch (Kletter et al., 2015; Porter et al., 2010). These resour- Defined glycan arrays such as those produced by the Consortium ces provide large amounts of raw data or automated presentation of for Functional Glycomics (CFG) and the National Center for binding motifs. However, the software have limited data V The Author(s) 2019. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 3536 Downloaded from https://academic.oup.com/bioinformatics/article-abstract/35/18/3536/5305635 by Ed 'DeepDyve' Gillespie user on 18 September 2019 GLAD: GLycan Array Dashboard 3537 visualization features (bar graphs and heatmaps) and are restrictive found in the GLAD documentation online along with video snippets with respect to usage (need login or special access), thus limiting and tutorials to guide the user on how to use the tools. In addition, we their use in visual analytics. To facilitate the data visualization and have put up an example demonstrating the ability to use GLAD to dif- visual comparison of glycan array data, we present here a new tool, ferentiate between Galectin-3, Galectin-3C and Galectin-7 online. GLAD (GLycan Array Dashboard), which is a browser-based tool to visualize and mine glycan array data. GLAD was developed using 3 Results and discussion a combination of JavaScript (JS) libraries including D3.js, jQuery, Lodash, Bootstrap and Select2js. As GLAD is mainly coded in JS, it GLAD is a unique tool offered to glycoscientists to assist in mining removes the need for users to upload their data to any server, and and comparing glycan array data in an easy and private manner. hence keeps user data and analysis private on their local site. This GLAD can be extremely useful in uncovering hidden relationships significantly reduces the need to maintain a high-performance ser- between array datasets, which was previously unexplored. The abil- ver. Furthermore, the data visualization includes not only bar graphs ity to visualize glycan array data, alongside glycan structure and the and heatmaps, but also calendar heatmaps, force graphs and correl- ability to filter the data based on user-defined criteria enables the ation maps. All visualizations are interactive and in SVG format for user to mine data in an efficient and thorough manner. In addition, the best user experience. GLAD also allows users to filter, sort and the selection files which are saved in the tool can be attached to normalize data to accentuate key data and binding relationships. manuscripts as Supplementary Information for reviewers and read- Overall, GLAD helps to make glycan array data analysis simpler ers to have a more interactive view of the data. The tool is constant- and more easily accessible to a larger community. ly being updated to add new features as requested by internal users of the NCFG, which provides glycan microarray assays as a service to the scientific community, and is open to suggestions from the gly- 2 Materials and methods coscience community. Additionally, newer data visualizations are constantly being investigated. In the future, we intend to host a data- The general workflow for GLAD is as follows (Fig. 1): 1. Input data base, which could feed data that is public into the GLAD interface as tab-delimited text files in the correct format. 2. Visualize and se- for visualization. This will allow users to analyze and compare their lect data for display using bar charts (Stage-1), or, alternatively add data along with publicly available data. the entire data file to the selections. 3. Save/Load the selections from multiple experiments. This allows the user to perform their analysis locally and resume at any time. Filter, Sort or Normalize the selec- Acknowledgements tions options allow the user to mine the data to analyze binding pat- terns. 4. Visualize the data selected from various experiments in one The authors would like to thank all members of the National Center for Functional Glycomics for their valuable feedback, especially Chao Gao, of the Stage-2 visualizations, i.e.: Lauren Byrd-Leotis, Tanya McKitrick, Jamie Heimburg-Molinaro and Alyssa Grouped Bar Charts- useful to compare small datasets. McQuillan. Heatmaps- useful to compare data for large numbers of samples across small numbers of glycans. Calendar Heatmaps- useful to compare data for large numbers Funding of glycans across multiple experiments. This work was supported by National Institutes of Health Grants Force Graphs- useful to visualize any network-like binding pat- P41GM103694 and U01GM125267. terns and identify common binding glycans among samples. Conflict of Interest: none declared. Correlation plots- useful to compare and quantify similarities or differences in binding patterns for a set of experiments, based on the generated Pearson Coefficient (r) values. References Bubble Box plots- allows users to get statistical values as box Agravat,S.B. et al. (2014) GlycoPattern: a web platform for glycan array min- plots and mean values with scatter-plot like data. ing. Bioinformatics, 30, 3417–3418. Cummings,R.D. and Pierce,J.M. (2014) The challenge and promise of glyco- The users can save and export the figures as SVG files. Users can mics. Chem. Biol., 21, 1–15. view the glycan structure in the Glycan Structure drawer at any time Heimburg-Molinaro,J. et al. (2011) Preparation and analysis of glycan micro- by hovering the mouse over data points. Full documentation can be arrays. Curr. Protoc. Protein Sci., 10, 12.10.1–12.10.29. Kletter,D. et al. (eds.) (2015) Glycoscience: Biology and Medicine. Springer, Tokyo, Japan, pp. 61–68. Liang,C.H. and Wu,C.Y. (2009) Glycan array: a powerful tool for glycomics studies. Exp. Rev. Proteomics, 6, 631–645. Porter,A. et al. (2010) A motif-based analysis of glycan array data to determine the specificities of glycan-binding proteins. Glycobiology, 20, 369–380. Seeberger,P.H. (2015) Monosaccharide diversity. In: Varki,A. et al. (eds.) Essentials of Glycobiology. Cold Spring Harbor, New York, pp. 19–30. Shelly,M.A. (1996) Exploratory data analysis: data visualization or torture? Infect. Control Hosp. Epidemiol., 17, 605–612. Varki,A. (2017) Biological roles of glycans. Glycobiology, 27, 3–49. Varki,A. et al. (eds.) (2015) Essentials of Glycobiology. Cold Spring Harbor, New York. pp. 77–88. Fig. 1. GLAD workflow showing how the user can input, manipulate and visu- Venkataraman,M. et al. (2015) Glycan array data management at consortium alize glycan array data for functional glycomics. Methods Mol. Biol., 1273, 181–190.

Journal

BioinformaticsOxford University Press

Published: Feb 1, 2019

There are no references for this article.