Scientific publication and news

Message boards : News : Scientific publication and news
Message board moderation

To post messages, you must log in.

AuthorMessage
damotbe
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Help desk expert

Send message
Joined: 23 Jul 19
Posts: 281
Credit: 457,739,361
RAC: 20,127
Message 1107 - Posted: 29 Sep 2020, 8:09:19 UTC

Dear Crunchers.

Thank you very much for your help.

I am pleased to announce the publication of our latest open access article describing EvoMol, our opensource molecule generator.

EvoMol: a flexible and interpretable evolutionary algorithm for unbiased de novo molecular generation
The objective of this work is to design a molecular generator capable of exploring known as well as unfamiliar areas of the chemical space.
Our method must be flexible to adapt to very different problems. Therefore, it has to be able to work with or without the influence of prior data and knowledge. Moreover, regardless of the success, it should be as interpretable as possible to allow for diagnosis and improvement.
We propose here a new open source generation method using an evolutionary algorithm to sequentially build molecular graphs. It is independent of starting data and can generate totally unseen compounds. To be able to search a large part of the chemical space, we define an original set of 7 generic mutations close to the atomic level.
Our method achieves excellent performances and even records on the QED, penalised logP, SAscore, CLscore as well as the set of goal-directed functions defined in GuacaMol. To demonstrate its flexibility, we tackle a very different objective issued from the organic molecular materials domain. We show that EvoMol can generate sets of optimised molecules having high energy HOMO or low energy LUMO, starting only from methane. We can also set constraints on a synthesizability score and structural features. Finally, the interpretability of EvoMol allows for the visualisation of its exploration process as a chemically relevant tree.


You can find for free the full article :
https://www.researchgate.net/publication/344326849_EvoMol_a_flexible_and_interpretable_evolutionary_algorithm_for_unbiased_de_novo_molecular_generation
or here :
https://jcheminf.biomedcentral.com/articles/10.1186/s13321-020-00458-z

You can help us a little bit more by visiting these pages to give us visibility and you can also share on your teams forums and/or social websites like tweeter.(@b_damota)

We have been working for some time now on the following article. It will deal in particular with the calculations you have made since the beginning of the project. The result will be an open access dataset. While we are writing this article we are also working on the next parts. Without divulging, I can only tell you that your calculations will help us to propose a unique tool that is particularly useful for chemists. We will of course keep you informed! Currently, two campaigns are in progress and do not concern the work mentioned above. The "nwchem long" units and the new "nwchem" with tasks with prefix "CL9" will also bring new results for articles probably end of 2021 or 2022.

Kindly
Benoit
ID: 1107 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
biodoc

Send message
Joined: 8 Dec 19
Posts: 4
Credit: 14,340,283
RAC: 10
Message 1110 - Posted: 29 Sep 2020, 15:16:44 UTC

Congratulations on the publication!
ID: 1110 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
bozz4science

Send message
Joined: 15 May 20
Posts: 5
Credit: 331,200
RAC: 3,945
Message 1113 - Posted: 30 Sep 2020, 9:40:25 UTC - in response to Message 1107.  

Well done! Thanks for making all this open to the public. I will have a look at this paper later today even though most of the technical terms in the paper abstract already intimidate me a lot. Anyways, I get the gist of what you are ultimately trying to achieve here and understanding even a little bit more detailed how that is implemented code-wise and some of the minuscule details of what techniques you apply there is just icing on the cake for me :) But I will try my hardest to comprehend the main aspects.

Benoit, just one question regarding the forward-looking statements relating to the subsequent articles/papers you mention to be scheduled for 2021/2022. Could you just indulge me what the "top-level" motivation or field is that you are trying to achieve with the development of this particular tool? Maybe you already touched upon this and it slipped my attention, but I would appreciate it anyway if you could tell us if this tool is specifically designed to target the molecule exploration in a particular field/industry or rather like a general-use tool.

So far, I couldn't contribute much as I was memory constrained and without a Linux partition on my system, I couldn't exploit the efficiency gains of those native based Linux apps. Hope to ramp it up a bit soon :)

Thanks! Again, great work!
ID: 1113 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
damotbe
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Help desk expert

Send message
Joined: 23 Jul 19
Posts: 281
Credit: 457,739,361
RAC: 20,127
Message 1114 - Posted: 30 Sep 2020, 11:49:27 UTC - in response to Message 1113.  

As a computer scientist, I try to develop methodologies, so things that are rather applicable everywhere and even outside of chemistry. For example, I have work related to genetics and neuroimaging. For applications in organic molecular chemistry (which is already a sub-domain), the methodology is applicable to medicine, and molecular materials. Thomas Cauchy, the theoretical chemist with whom I work is in a laboratory whose specialty is molecular materials, organic photovoltaics, organic LEDs, and sometimes even compounds suitable for phototherapy. For 2021/2022, we are going to work on methodological tools, then implementations to propose more stable and more relevant molecules. This method will be applicable to all applications where it is a question of finding new molecules (de novo molecular generation).

Well done! Thanks for making all this open to the public. I will have a look at this paper later today even though most of the technical terms in the paper abstract already intimidate me a lot. Anyways, I get the gist of what you are ultimately trying to achieve here and understanding even a little bit more detailed how that is implemented code-wise and some of the minuscule details of what techniques you apply there is just icing on the cake for me :) But I will try my hardest to comprehend the main aspects.

Benoit, just one question regarding the forward-looking statements relating to the subsequent articles/papers you mention to be scheduled for 2021/2022. Could you just indulge me what the "top-level" motivation or field is that you are trying to achieve with the development of this particular tool? Maybe you already touched upon this and it slipped my attention, but I would appreciate it anyway if you could tell us if this tool is specifically designed to target the molecule exploration in a particular field/industry or rather like a general-use tool.

So far, I couldn't contribute much as I was memory constrained and without a Linux partition on my system, I couldn't exploit the efficiency gains of those native based Linux apps. Hope to ramp it up a bit soon :)

Thanks! Again, great work!
ID: 1114 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
bozz4science

Send message
Joined: 15 May 20
Posts: 5
Credit: 331,200
RAC: 3,945
Message 1124 - Posted: 3 Oct 2020, 10:55:19 UTC - in response to Message 1114.  

Hey Benoit,
that was a very interesting read! Merci beaucoup pour ta vite réponse très informative! I appreciate it very much!

Well, your answer covers just about any field that is currently driving innovation in these science domains. I am thrilled that there are so many potential use cases where this tool could be used to help spark further innovation and explorative science. Many of which I haven't even considered before your response. Very exciting future!

Considering the second step you are talking about to develop a methodology to filter out the most interesting and relevant molecules that were generated using the de novo molecular generation algorithm, will this be based on machine learning/statistical optimization?

The only thing that I was hesitant about first, was that I read somewhere (I believe it was this project's "home" page) that QuChemPedIA was trying to develop an AI system that could potentially develop and optimise yet unknown synthetic molecules with any property you as a researcher desire. I feared that this could potentially be misused and exploited for like let's say weapons research or the like. But in the end I guess, making this tool that is the very product of your research openly available to the public frankly is the right way to proceed, as any innovative solution can eventually be exploited or abused for malicious purpose. And in the end you will still need to require tremendous domain knowledge (in the respective field as well as chemist know-how) to being able to properly handle the tool and get the results you want as well as massive computing power. So there's that. In the end, I enjoyed reading your answer and all the mentioned use cases that could create amazing technical solutions for many of mankind's problem. I love that QuChemPedIA is doing basic methodological science first to create this general purpose tool that can then in return ultimately advance any and every research field that handles "molecules" in some way.
ID: 1124 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
A_Grumpy_Turtle

Send message
Joined: 10 Sep 20
Posts: 1
Credit: 86,400
RAC: 0
Message 1126 - Posted: 3 Oct 2020, 17:06:26 UTC

Nice work!
ID: 1126 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : News : Scientific publication and news

©2021 Benoit DA MOTA - LERIA, University of Angers, France