paper related to modeling of chemical reactions

[UnixJunkie]

Hello,

I cannot find your e-mail, so I am using this issue.

You might be interested in this paper:
"Structure–reactivity modeling using mixture-based representation of chemical reactions"
https://link.springer.com/article/10.1007%2Fs10822-017-0044-3

I saw your presentation on prediction of chemical reactions using DNN in Kumamoto recently.

The encoding presented in this paper is better than ECFP4 I think.
"this approach doesn't need any explicit information about reaction center and therefore
no atom-to-atom mapping is required"

Regards,
Francois.