I know that is a very general question but I would like to start a ML project in Python to predict some chemical properties with a large set of experimental data. The compounds I would like to study are transition metal complexes so I have difficult generating SMILES, for example. I would like to know if there are any books (or other sources) on which to learn well what descriptors are, which ones to use, how to apply them, which databases to use, how to choose a regression model or how to faithfully convert coordination bonds of complexes.
$\begingroup$
$\endgroup$
4
-
$\begingroup$ Fingerprints are usually made for drug-like molecules, I don't know how effective they would be for metal complexes. You can certainly try graph convolutional neural networks which read the connectivity of a molecule (as a graph), e.g. with DeepChem $\endgroup$– S R MaitiFeb 16, 2022 at 10:46
-
$\begingroup$ One possible start is to try to reproduce a paper that you find interesting. Especially, if you have no experience in the field, it helps to find your own errors, too, and often have specific codes published, so you can compare your solution and learn. $\endgroup$– GregFeb 16, 2022 at 12:01
-
$\begingroup$ +1 welcome to our new community and thank you for contributing your question here! We hope to see much more of you in the future!!! How did you find this site? $\endgroup$– Nike DattaniFeb 16, 2022 at 20:22
-
$\begingroup$ @NikeDattani Thank you so much, I already like this community, I found it by accident by googling. $\endgroup$– vorsamaqoyFeb 17, 2022 at 10:12
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Admittedly, there are tons of materials on the chemistry + machine learning topics. Let me give one:
An introductory text I find useful is Machine Learning in Chemistry from Janet and Kulik in ACS in Focus series. If you check the papers of the Kulik group at MIT you will find much insightful information on which descriptors of coordination compounds work and ties up nicely with the above-mentioned tutorial by the same authors.
