New chemiscope JOSS paper

paper/paper-2020/paper-2020.md

@@ -0,0 +1,120 @@
+---
+title: 'Chemiscope: interactive structure-property explorer for materials and molecules'
+tags:
+    - TypeScript
+    - JavaScript
+    - chemistry
+    - material science
+    - machine learning
+authors:
+    - name: Guillaume Fraux
+      orcid: 0000-0003-4824-6512
+      affiliation: 1
+    - name: Rose K. Cersonsky
+      orcid: 0000-0003-4515-3441
+      affiliation: 1
+    - name: Michele Ceriotti
+      orcid: 0000-0003-2571-2832
+      affiliation: 1
+affiliations:
+    - name: Laboratory of Computational Science and Modeling, IMX, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland
+      index: 1
+date: 30 January 2020
+bibliography: paper.bib
+---
+
+# Summary
+
+The number of materials or molecules that can be created by combining different
+chemical elements in various proportions and spatial arrangements is enormous.
+Computational chemistry can be used to generate databases containing billions of
+potential structures [@Ruddigkeit2012], and predict some of the associated
+properties [@Montavon2013; @Ramakrishnan2014]. Unfortunately, the very large
+number of structures makes exploring such database — to understand
+structure-property relations or find the _best_ structure for a given
+application — a daunting task. In recent years, multiple molecular
+_representations_ [@Behler2007; @Bartok2013; @Willatt2019] have been developed
+to compute structural similarities between materials or molecules, incorporating
+physically-relevant information and symmetries. The features associated with
+these representations can be used for unsupervised machine learning
+applications, such as clustering or classification of the different structures,
+and high-throughput screening of database for specific properties [@Maier2007;
+@De2017; @Hautier2019]. Unfortunately, the dimensionality of these features (as
+well as most of other descriptors used in chemical and materials informatics) is
+very high, which makes the resulting classifications, clustering or mapping very
+hard to visualize. Dimensionality reduction algorithms [@Schlkopf1998;
+@Ceriotti2011; @McInnes2018] can reduce the number of relevant dimensions to a
+handful, creating 2D or 3D maps of the full database.
+
+![The Qm7b database [@Montavon2013] visualized with chemiscope](screenshot.png)
+
+Chemiscope is a graphical tool for the interactive exploration of materials and
+molecular databases, correlating local and global structural descriptors with
+the physical properties of the different systems. The interface consists of
+two panels. The left panel displays a 2D or 3D scatter plot, in which each
+point corresponds to a chemical entity. The axes, color, and style of each point
+can be set to represent a property or a structural descriptor to visualize
+structure-property relations directly. Structural descriptors are not computed
+directly by chemiscope, but must be obtained from one of the many codes
+implementing general-purpose atomic representation [@librascal; @QUIP] or more specialized descriptors. Since the most common
+descriptors can be very high dimensional, it can be convenient to apply a
+dimensionality reduction algorithm that maps them to a lower-dimensional space
+for easier visualization. For example the sketch-map algorithm [@Ceriotti2011]
+was used with the Smooth Overlap of Atomic Positions representation [@Bartok2013] to
+generate the visualization in Figure 1. The right panel displays the
+three-dimensional structure of the chemical entities, possibly including
+periodic repetition for crystals. Visualizing the chemical structure can help
+in finding an intuitive rationalization of the layout of the dataset and the
+structure-property relations.
+
+Whereas similar tools [@Gong2013; @Gutlein2014; @Probst2017; @ISV] only allow
+visualizing maps and structures in which each data point corresponds to a
+molecule, or a crystal structure, a distinctive feature of chemiscope is the
+possibility of visualizing maps in which points correspond to atom-centred
+environments. This is useful, for instance, to rationalize the relationship
+between structure and atomic properties such as nuclear chemical shieldings
+(Figure 2). This is also useful as a diagnostic tool for the many
+machine-learning schemes that decompose properties into atom-centred
+contributions [@Behler2007; @Bartok2010].
+
+![Database of chemical shieldings [@Paruzzo2018] in chemiscope demonstrating the use of a 3D plot and highlighting of atomic environments](./screenshot-3d.png)
+
+Chemiscope took strong inspiration from a previous similar graphical software,
+the interactive sketch-map visualizer [@ISV]. This previous software was used in
+multiple research publication, related to the exploration of large-scale
+databases, and the mapping of structure-property relationships [@De2016;
+@De2017; @Musil2018].
+
+# Implementation
+
+Chemiscope is implemented using the web platform: HTML5, CSS and WebGL to
+display graphical elements, and TypeScript (compiled to JavaScript) for
+interactivity. It uses [Plotly.js](https://plot.ly/javascript/) to render and
+animate 2D and 3D plots; and the JavaScript version of [Jmol](http://jmol.org/)
+to display atomic structures. The visualization is fast enough to be used with
+datasets containing up to a million points, reacting to user input within a few
+hundred milliseconds in the default 2D mode. More elaborate visualizations are
+slower, while still handling 100k points easily.
+
+The use of web technologies makes chemiscope usable from different operating
+systems without the need to develop, maintain and package the code for each
+operating system. It also means that we can provide an online service at
+http://chemiscope.org that allows users to visualize their own dataset without any
+local installation. Chemiscope is implemented as a library of re-usable
+components linked together via callbacks. This makes it easy to modify the
+default interface to generate more elaborate visualizations, for example,
+displaying multiple maps generated with different parameters of a dimensionality
+reduction algorithm. Chemiscope can also be distributed in a standalone mode,
+where the code and a predefined dataset are merged together as a single HTML
+file. This standalone mode is useful for archival purposes, for example as
+supplementary information for a published article and for use in corporate
+environments with sensitive datasets.
+
+# Acknowledgements
+
+The development of chemiscope have been funded by the [NCCR
+MARVEL](http://nccr-marvel.ch/), the [MAX](http://max-centre.eu/) European
+centre of excellence, and the European Research Council (Horizon 2020 grant
+agreement no. 677013-HBMAP).
+
+# References

paper/paper-2020/paper.bib

@@ -0,0 +1,273 @@
+@article{Ceriotti2011,
+  doi = {10.1073/pnas.1108486108},
+  url = {https://doi.org/10.1073/pnas.1108486108},
+  year = {2011},
+  month = {07},
+  publisher = {Proceedings of the National Academy of Sciences},
+  volume = {108},
+  number = {32},
+  pages = {13023--13028},
+  author = {Michele Ceriotti and Gareth A. Tribello and Michele Parrinello},
+  title = {Simplifying the representation of complex free-energy landscapes using sketch-map},
+  journal = {Proceedings of the National Academy of Sciences}
+}
+
+@article{Bartok2013,
+  doi = {10.1103/physrevb.87.184115},
+  url = {https://doi.org/10.1103/physrevb.87.184115},
+  year = {2013},
+  month = {05},
+  publisher = {American Physical Society ({APS})},
+  volume = {87},
+  number = {18},
+  author = {Albert P. Bart{\'{o}}k and Risi Kondor and G{\'{a}}bor Cs{\'{a}}nyi},
+  title = {On representing chemical environments},
+  journal = {Physical Review B}
+}
+
+@article{Montavon2013,
+  doi = {10.1088/1367-2630/15/9/095003},
+  url = {https://doi.org/10.1088/1367-2630/15/9/095003},
+  year = {2013},
+  month = {09},
+  publisher = {{IOP} Publishing},
+  volume = {15},
+  number = {9},
+  pages = {095003},
+  author = {Grégoire Montavon and Matthias Rupp and Vivekanand Gobre and Alvaro Vazquez-Mayagoitia and Katja Hansen and Alexandre Tkatchenko and Klaus-Robert M\"{u}ller and O Anatole von Lilienfeld},
+  title = {Machine learning of molecular electronic properties in chemical compound space},
+  journal = {New Journal of Physics}
+}
+
+@article{Gutlein2014,
+  doi = {10.1186/s13321-014-0041-7},
+  url = {https://doi.org/10.1186/s13321-014-0041-7},
+  year = {2014},
+  month = sep,
+  publisher = {Springer Science and Business Media {LLC}},
+  volume = {6},
+  number = {1},
+  author = {Martin G\"{u}tlein and Andreas Karwath and Stefan Kramer},
+  title = {{CheS}-Mapper 2.0 for visual validation of (Q){SAR} models},
+  journal = {Journal of Cheminformatics}
+}
+
+@article{Probst2017,
+  doi = {10.1093/bioinformatics/btx760},
+  url = {https://doi.org/10.1093/bioinformatics/btx760},
+  year = {2017},
+  month = {10},
+  publisher = {Oxford University Press ({OUP})},
+  volume = {34},
+  number = {8},
+  pages = {1433--1435},
+  author = {Daniel Probst and Jean-Louis Reymond},
+  editor = {Jonathan Wren},
+  title = {{FUn}: a framework for interactive visualizations of large,  high-dimensional datasets on the web},
+  journal = {Bioinformatics}
+}
+
+@article{Gong2013,
+  doi = {10.1093/bioinformatics/btt270},
+  url = {https://doi.org/10.1093/bioinformatics/btt270},
+  year = {2013},
+  month = {05},
+  publisher = {Oxford University Press ({OUP})},
+  volume = {29},
+  number = {14},
+  pages = {1827--1829},
+  author = {Jiayu Gong and Chaoqian Cai and Xiaofeng Liu and Xin Ku and Hualiang Jiang and Daqi Gao and Honglin Li},
+  title = {{ChemMapper}: a versatile web server for exploring pharmacology and chemical structure association based on molecular 3D similarity method},
+  journal = {Bioinformatics}
+}
+
+@article{Paruzzo2018,
+  doi = {10.1038/s41467-018-06972-x},
+  url = {https://doi.org/10.1038/s41467-018-06972-x},
+  year = {2018},
+  month = oct,
+  publisher = {Springer Science and Business Media {LLC}},
+  volume = {9},
+  number = {1},
+  author = {Federico M. Paruzzo and Albert Hofstetter and Félix Musil and Sandip De and Michele Ceriotti and Lyndon Emsley},
+  title = {Chemical shifts in molecular solids by machine learning},
+  journal = {Nature Communications}
+}
+
+@software{ISV,
+  author = {De, Sandip and Ceriotti, Michele},
+  title = {Interactive Sketchmap Visualizer},
+  publisher = {Zenodo},
+  year = {2019},
+  version = {1.0.0},
+  doi = {10.5281/zenodo.3541831},
+  url = {https://doi.org/10.5281/zenodo.3541831}
+}
+
+@article{De2016,
+  doi = {10.1039/c6cp00415f},
+  url = {https://doi.org/10.1039/c6cp00415f},
+  year = {2016},
+  publisher = {Royal Society of Chemistry ({RSC})},
+  volume = {18},
+  number = {20},
+  pages = {13754--13769},
+  author = {Sandip De and Albert P. Bart{\'{o}}k and G{\'{a}}bor Cs{\'{a}}nyi and Michele Ceriotti},
+  title = {Comparing molecules and solids across structural and alchemical space},
+  journal = {Physical Chemistry Chemical Physics}
+}
+
+@article{De2017,
+  doi = {10.1186/s13321-017-0192-4},
+  url = {https://doi.org/10.1186/s13321-017-0192-4},
+  year = {2017},
+  month = {02},
+  publisher = {Springer Science and Business Media {LLC}},
+  volume = {9},
+  number = {1},
+  author = {Sandip De and Félix Musil and Teresa Ingram and Carsten Baldauf and Michele Ceriotti},
+  title = {Mapping and classifying molecules from a high-throughput structural database},
+  journal = {Journal of Cheminformatics}
+}
+
+@article{Musil2018,
+  doi = {10.1039/c7sc04665k},
+  url = {https://doi.org/10.1039/c7sc04665k},
+  year = {2018},
+  publisher = {Royal Society of Chemistry ({RSC})},
+  volume = {9},
+  number = {5},
+  pages = {1289--1300},
+  author = {Félix Musil and Sandip De and Jack Yang and Joshua E. Campbell and Graeme M. Day and Michele Ceriotti},
+  title = {Machine learning for the structure-energy-property landscapes of molecular crystals},
+  journal = {Chemical Science}
+}
+
+@article{Hautier2019,
+  doi = {10.1016/j.commatsci.2019.02.040},
+  url = {https://doi.org/10.1016/j.commatsci.2019.02.040},
+  year = {2019},
+  month = {06},
+  publisher = {Elsevier {BV}},
+  volume = {163},
+  pages = {108--116},
+  author = {Geoffroy Hautier},
+  title = {Finding the needle in the haystack: Materials discovery and design through computational ab initio high-throughput screening},
+  journal = {Computational Materials Science}
+}
+
+@article{Willatt2019,
+  doi = {10.1063/1.5090481},
+  url = {https://doi.org/10.1063/1.5090481},
+  year = {2019},
+  month = {04},
+  publisher = {{AIP} Publishing},
+  volume = {150},
+  number = {15},
+  pages = {154110},
+  author = {Michael J. Willatt and F{\'{e}}lix Musil and Michele Ceriotti},
+  title = {Atom-density representations for machine learning},
+  journal = {The Journal of Chemical Physics}
+}
+
+@article{Behler2007,
+  doi = {10.1103/physrevlett.98.146401},
+  url = {https://doi.org/10.1103/physrevlett.98.146401},
+  year = {2007},
+  month = {04},
+  publisher = {American Physical Society ({APS})},
+  volume = {98},
+  number = {14},
+  author = {J\"{o}rg Behler and Michele Parrinello},
+  title = {Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces},
+  journal = {Physical Review Letters}
+}
+
+@article{Ruddigkeit2012,
+  doi = {10.1021/ci300415d},
+  url = {https://doi.org/10.1021/ci300415d},
+  year = {2012},
+  month = {11},
+  publisher = {American Chemical Society ({ACS})},
+  volume = {52},
+  number = {11},
+  pages = {2864--2875},
+  author = {Lars Ruddigkeit and Ruud van Deursen and Lorenz C. Blum and Jean-Louis Reymond},
+  title = {Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database {GDB}-17},
+  journal = {Journal of Chemical Information and Modeling}
+}
+
+@article{Ramakrishnan2014,
+  doi = {10.1038/sdata.2014.22},
+  url = {https://doi.org/10.1038/sdata.2014.22},
+  year = {2014},
+  month = {08},
+  publisher = {Springer Science and Business Media {LLC}},
+  volume = {1},
+  number = {1},
+  author = {Raghunathan Ramakrishnan and Pavlo O. Dral and Matthias Rupp and O. Anatole von Lilienfeld},
+  title = {Quantum chemistry structures and properties of 134 kilo molecules},
+  journal = {Scientific Data}
+}
+
+@article{Bartok2010,
+  doi = {10.1103/physrevlett.104.136403},
+  url = {https://doi.org/10.1103/physrevlett.104.136403},
+  year = {2010},
+  month = {04},
+  publisher = {American Physical Society ({APS})},
+  volume = {104},
+  number = {13},
+  author = {Albert P. Bart{\'{o}}k and Mike C. Payne and Risi Kondor and G{\'{a}}bor Cs{\'{a}}nyi},
+  title = {Gaussian Approximation Potentials: The Accuracy of Quantum Mechanics,  without the Electrons},
+  journal = {Physical Review Letters}
+}
+
+@article{Schlkopf1998,
+  doi = {10.1162/089976698300017467},
+  url = {https://doi.org/10.1162/089976698300017467},
+  year = {1998},
+  month = {08},
+  publisher = {{MIT} Press - Journals},
+  volume = {10},
+  number = {5},
+  pages = {1299--1319},
+  author = {Bernhard Sch\"{o}lkopf and Alexander Smola and Klaus-Robert M\"{u}ller},
+  title = {Nonlinear Component Analysis as a Kernel Eigenvalue Problem},
+  journal = {Neural Computation}
+}
+
+@article{Maier2007,
+  doi = {10.1002/anie.200603675},
+  url = {https://doi.org/10.1002/anie.200603675},
+  year = {2007},
+  month = aug,
+  publisher = {Wiley},
+  volume = {46},
+  number = {32},
+  pages = {6016--6067},
+  author = {Wilhelm{\hspace{0.25em}}F. Maier and Klaus St\"{o}we and Simone Sieg},
+  title = {Combinatorial and High-Throughput Materials Science},
+  journal = {Angewandte Chemie International Edition}
+}
+
+@article{McInnes2018,
+    title={UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction},
+    author={Leland McInnes and John Healy and James Melville},
+    year={2018},
+    eprint={1802.03426},
+}
+
+@online{librascal,
+    author = {librascal},
+    title = {},
+    date = {},
+    url = {https://github.com/lab-cosmo/librascal}
+}
+
+@online{QUIP,
+    author = {QUIP},
+    title = {},
+    date = {},
+    url = {http://libatoms.github.io/QUIP/}
+}