MEXPLORER 1.0.0-rc.0 &ndash; A mechanism explorer for analysis and visualization of chemical reaction pathways based on graph theory

Sander, Rolf

doi:https://doi.org/10.5194/egusphere-2023-1577

Preprints

https://doi.org/10.5194/egusphere-2023-1577

Preprints

02 Aug 2023

| 02 Aug 2023

Status: this preprint is open for discussion.

MEXPLORER 1.0.0-rc.0 – A mechanism explorer for analysis and visualization of chemical reaction pathways based on graph theory

Rolf Sander

Abstract. The open-source software MEXPLORER 1.0.0-rc.0¹ is presented here. The program can be used to analyze, reduce, and visualize complex chemical reaction mechanisms. The mathematics behind the tool is based on graph theory: chemical species are represented as vertices, and reactions as edges. MEXPLORER is a community model published under the GNU General Public License.

¹The name of this version indicates that it is a release candidate used for the interactive discussion in GMDD. If necessary, bug fixes can still be made. The release of the final version MEXPLORER-1.0.0 is planned together with the final paper in GMD.

Received: 10 Jul 2023 – Discussion started: 02 Aug 2023

Download & links

Rolf Sander

Status: open (until 27 Sep 2023)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2023-1577', Anonymous Referee #1, 14 Sep 2023 reply
General comments
The manuscript by R. Sander presents an open-source code for the management of complex reaction mechanisms, employing a graph-based representation to enable the utilization of several tools from Graph Theory. The capabilities of the software are clearly highlighted, mentioning the different functionalities that MEXPLORER provides to visualize, filter, build and evaluate complex reaction mechanisms. Nonetheless, some of these aspects are not too contextualized, and I think that the manuscript could be more complete by providing additional details and references beyond this direct showcase of functionalities: in their absence, the article seems, in some parts, more like a user manual for the program. Consequently, I would suggest some major revisions before recommending publication.
Specific comments (major revisions)

Line 25. Other possible domains of application of MEXPLORER are proposed (“e.g., chemical engineering and marine chemistry”), but no references of specific complex reaction mechanisms in these fields are given. Additionally, I believe that computational chemistry should also be mentioned, since many recent efforts are tackling, precisely, the automated generation of very complex reaction mechanisms from DFT calculations and, consequently, its visualization and treatment. Relevant and recent examples could be the work from Prof. Maeda (GRRM: [S. Maeda, Y. Harabuchi, H. Hayashi, T. Mita, Annu. Rev. Phys. Chem., 2023, 74], [Y. Sumiya, Y. Harabuchi, Y. Nagata, S. Maeda, JACS Au, 2022, 2, 1181-1188) or Prof. Martínez-Núñez (AutoMeKin: [E. Martínez-Nuñez, G.L. Barnes, D.R. Glowacki, S. Kopec, D. Pelaez, A. Rodriguez, R. Rodriguez-Fernandez, R.J. Shannon, J.J.P. Stewart, P.G. Tahoces, S.A. Vazquez, J. Comput. Chem., 2021, 42, 2036], [D. Garay-Ruiz, M. Álvarez-Moreno, C. Bo, E. Martínez-Núñez, ACS Phys. Chem. Au, 2022, 2, 3, 225-336]), among others.

Lines 27 – 31. The choice of representing reactions as edges between reagents and products is justified as “this produces plots with much less visual clutter”. While this is correct, to me it seems a bit of an oversimplification, reducing graphs to a mere visualization tool. A more “conceptual” approach would enrich the discussion, introducing, for example, how this reduction in the size of the graph (against bipartite graphs as proposed by Silva et al.) may also imply faster processing, simplify path finding algorithms, and so on.

Line 65. The OIC (overall interaction coefficient) is introduced, but it is not explained. The corresponding equation and/or a brief summary of how this value can be obtained should be included.

Line 71. The affirmation “Alternative representations (...) can be created with additional Python functions, if needed” does not seem necessary, in the case that it is referred to the users being able to write custom functions and expand the code. If it refers instead to specific existing functions in MEXPLORER that have not been discussed along the section, a brief enumeration of these would be of interest.

Lines 80 – 85. Regarding the function to merge parallel edges, another possible simplification would be to take the “best” reaction of the bunch as the representative of the group (e.g. the one with the largest rate constant), so a simpler graph can be obtained where every edge is conceptually a single reaction, with a single set of names and parameters. Is this available in some way/is there a reason not to do it?

Line 86. Same as point 3.

Line 102. Having a filtering function for the specific JAMOC mechanism seems a bit specific. Adding some more context on the importance of this could make it clearer.

Line 139 (max-flow problem section). While the available algorithms are properly listed and referenced, given the importance and difficulty of the general problem of locating important reaction pathways in any complex mechanism, a more thorough discussion would be quite interesting. For instance, introducing the key differences between the three available algorithms, as well as an idea on the timings that they take for increasingly complex networks. Enriching this part of the discussion would reinforce the manuscript as a whole.

Lines 167 – 169. The wording “movies with changing arrow widths, visualizing how important pathways may change through the diurnal cycle” hints more at a very specific example than to a general feature of the code such as the time evolution of reaction networks and the corresponding representation. I would suggest rephrasing the paragraph to first comment on the idea of network evolution, then on the possibility of plotting movies, and only then to the specific day-cycle idea.

Technical comments:
Figure sizes are quite inconsistent. While it is understandable that this kind of automatically generated complex graphs might have varying shapes and sizes, a general consistency would be desirable. For instance, Figures 4 and 5 could be reduced so the general node and font sizes match Figures 2 and 3.

Some of the node colors are a bit dark to allow the text to be comfortably read, especially in smaller sizes. Figure 3 is particularly hard to read, as it is composed by green and turquoise nodes that encounter this problem. Using paler shades of these two colors will make the final result much more readable.

Reply
Citation: https://doi.org/10.5194/egusphere-2023-1577-RC1

Rolf Sander

Viewed

Total article views: 302 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
203	91	8	302	5	4

HTML: 203
PDF: 91
XML: 8
Total: 302
BibTeX: 5
EndNote: 4

Views and downloads (calculated since 02 Aug 2023)

Month	HTML	PDF	XML	Total
Aug 2023	164	80	5	249
Sep 2023	39	11	3	53

Cumulative views and downloads (calculated since 02 Aug 2023)

Month	HTML	PDF	XML	Total
Aug 2023	164	80	5	249
Sep 2023	39	11	3	53

Viewed (geographical distribution)

Total article views: 296 (including HTML, PDF, and XML) Thereof 296 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 22 Sep 2023

Short summary

The open-source software MEXPLORER 1.0.0-rc.0 is presented here. The program can be used to analyze, reduce, and visualize complex chemical reaction mechanisms. The mathematics behind the tool is based on graph theory: chemical species are represented as vertices, and reactions as edges. MEXPLORER is a community model published under the GNU General Public License.


Total:	0
HTML:	0
PDF:	0
XML:	0