Should there be a Chemistry Domain/Organization?

Should there be a chemistry domain/organization? I think this would have very few users… But, there are a few packages of this nature (maybe 5-6 now?). I am a chemist by training so this interests me, but I don’t know it would have a lot of lift or if that really matters or not.

7 Likes

I don’t see how a chemistry org could hurt.
Sometimes it’s like a self-fulfilling prophecy, having an organization might encourage more Chem ppl to use Julia…
Check out: Task Views on Chemometrics
Would your org scope be Chemistry broadly defined (beyond Chemometrics)?

2 Likes

I obtained my Ph.D doing Chemometrics :). (self plug: GitHub - caseykneale/ChemometricsTools.jl: A collection of tools for chemometrics and machine learning written in Julia.)

Yea it could extend to molecular dynamics, thermomodelling, chemoinformatics, etc. Even things like data mining specific tools, reaction product forecasts, kinetics, who knows I guess :).

2 Likes

I would fully agree with that. Julia is definitely underrepresented in chemistry as of now. As you mention there are a few packages out there, see for example this list:

Personally, as one of the core developers of https://dftk.org and member of the JuliaMolSim organisation, I appreciate any motion towards in the direction of establishing Julia! Thanks for hitting this off @anon92994695!

Regarding the question of scope, I’d include quantum-chemical simulation methods (molecular electronic structure theory, quantum dynamics, etc.), but also methods in materials science (e.g. plane-wave DFT, tight-binding, atomistic models, …).

5 Likes

@mfh I wasn’t aware of your packages/org! Awesome! If this thing kicks off, would you like to be an organization owner? I can happily contribute especially from the informatics and chemometrics end. I do some thermo and kinetics stuff as well, but @longemen3000 is our resident thermo person I think.

will be registerring a SMILES package soon: https://github.com/caseykneale/OpenSMILES.jl

4 Likes

Sure, I would be happy to contribute in any way I can. I’m mostly a quantum-chemical simulation / electronic structure guy, so from your perspective a bit more on the first-principles, low-level end I guess :smile:. Please keep me in the loop!

3 Likes

okay the more digging I do the more I find there are a lot of chemistry packages in Julia

just found this one: GitHub - JuliaMolSim/Molly.jl: Molecular simulation in Julia

someone says there is a ChemE org? We can definitely pool these things together into a bigger thing for discoverability. Even if the big thing is just a directory to all the other things. I added a slack channel #chemistry feel free to join.

found this too: Julia.jl/Chemistry.md at master · svaksha/Julia.jl · GitHub !!

@dilumaluthge not sure if you are the correct person to ask but you seem to know this stuff or at least know who knows this stuff. what has to be done for there to be a chemistry Org? can we get advice from the Bio org on how they organize things? There are a lot of subsets here many of them under active development.

I think in this case doing some organization could help everyones disparate efforts.

5 Likes

Molly.jl is quite early stage at the minute but once it’s more mature I would not be against moving it to a new or existing chemistry org.

I’m interested to see how this develops! I have a chemistry background but don’t do it day-to-day anymore.

3 Likes

this is a biased opinion of what packages of other organizations fit in a made up definition of Chemistry, and interesting related packages with no particular order. i’m including julia-only packages.
BioJulia:
https://github.com/BioJulia/BioSymbols.jl: compounds names with associated types
JuliaPhysics:
https://github.com/JuliaPhysics/PhysicalConstants.jl (we need the R constant at least)
https://github.com/JuliaPhysics/Lattices.jl
https://github.com/JuliaPhysics/PeriodicTable.jl (!!!)
JuliaMolSim
https://github.com/JuliaMolSim/JuLIP.jl (check differences with Lattices.jl later)
https://github.com/JuliaMolSim/DFTK.jl (not my area, but very interesting)

3 Likes

Great finds!
I’m working on a fork for updating svashka’s list
https://github.com/caseykneale/Julia.jl/blob/master/Chemistry.md
reorganizing and adding onto it

But c’mon periodic table.jl? That’s chemistry not physics :stuck_out_tongue:

1 Like

What about a package/DataFrame/Database of thermo-physical data (experimental data), in combination with thermo-physical correlations? Does such a thing exist? Or possibly a Julia interface to some free database?

3 Likes

Thanks @anon92994695 , I am neither a Physicist nor a Chemist so PRs (now merged) from the domain experts are more than welcome!

2 Likes

Your list was really good how it originally was! Now it suits me but hopefully someone else will chime in so it suits everyone else :). Thank you svaksha for getting the groundwork laid for us :slight_smile:

1 Like

JuLIP.jl provides all sort of infrastructure for molecular simulation, similar in a way to ASE but nowhere near as complete and polished. It is not (yet) intended for production use, but rather for prototyping and to provide an interface to ASE. Providing mechanisms to describe different lattice structures is just one of the tools we use to construct material configurations for testing and for some basic modelling applications.

As a typical use case, we are developing new interatomic potentials (and parameter fitting algorithms) for various materials and macro-molecules, but in practise no modeller will switch to Julia to use these potentials so we provide an interface to ASE (and hopefully soon to LAMMPS). But developing these potentials in Python is a pain because we need the speed. (And I refuse to write C, C++ or Fortran)

Much of our infrastructure is quite experimental since we are just a small group and are interested mostly in theory and methodology. Maybe it will grow eventually into more broadly usable packages, maybe not. We are very much open to participation.

2 Likes

Great work. For me a few of the categories belong merged. For example Molecular Modelling and Quantum Chemistry I would merge under a category “Simulation Methods” and also include “Computational Tools” as a subcategory. This is also apparent from the fact, that there are duplicate entries at the moment. If there are no objections, I’ll file a PR.

Also we should include in this category:
https://github.com/vonDonnerstein/QuantumLab.jl

2 Likes

mfh do as you wish :). I’m not seasoned in the quantum end of things, and gosh I just rushed through adding what I could find. Team effort here will make this a really awesome list and from that list we can look at statistics/concepts and map out what we should be doing or if anything else ought to happen(IE: make an org or whatever)

2 Likes

Please file a PR. FWIW, I file packages under various catagory
top-levels and rely on the wikipedia CATs for help with topics that I
am not trained in. WP isnt perfect but they are a nice resource to
turn to when the first point of info is missing - read, package
authors description of what their package does.
Some authors simply dump code and the hardest part is reading the code
and trying to figure out what it does (with lots of error margins
thrown in :-P) by searching, reading, looking at their profile to
cross-check the domain, etc. I am learning everyday.

॥ SVAKSHA ॥ स्वक्ष ॥ BIO: http://svaksha.com/pages/BIO ॥ PUB:
http://svaksha.com/pages/PUB ॥ Git:: https://github.com/svaksha,
https://gitlab.com/svaksha ॥ T:: https://twitter.com/svaksha

2 Likes

Here it is: https://github.com/svaksha/Julia.jl/pull/180. Happy to discuss all the details. I mainly condensed a few categories and added a few packages of my own. I’m mostly in the physical chemistry / quantum chemistry field, so there I did most restructuring … the other things I’m not an expert in. Two packages don’t fit that well in my opinion and I put them into Uncategorized.

2 Likes

For everyone interested, there is a category named Quantum. It’s out there for quantum physics, chemistry, informatics, biology(?).

5 Likes

Hello dear all,

I have taught organic chemistry and physical organic chemistry during my career. My research activities focused on emission spectroscopy of species such as C2*, OH*, Na-Ar*, etc. That is to say, transitions of an electronically excited state, which simultaneously involves vibrational and rotational transitions (as in fluorescence).

Unless this is solved at the present time, one of the major problems in organic chemistry concerns the representation of organic molecules, the representation of reactions in a “clean” way and the representation of mechanisms.
Chemists waste a lot of time representing these components. Including for their publications.

Could we imagine a tablet (connected in one way or another to a computer) on which the chemist draws by hand (with a pen) his molecules and reactions. The computer could recognize the patterns (AI…) to restructure them neatly for lectures or publications.

One step further…But it may be a dream…
In a lecture, a professor speaks and, depending on the circumstances, verbally quotes a molecule and engages it in a reaction. The reaction is displayed “cleanly” and the students receive the information on their computer, their tablet. This would bring incredible flexibility to teaching and lectures. Of course, students could make some trials to learn with high performances.

Is this possible or not? With Julia, of course.

2 Likes