This blog have a collection of posts centered around useful information and practical examples of how to do various cheminformatic tasks going more and more into deep learning and AI examples. The blog is a continuation of the blog I wrote while I worked as an independent consultant with the company Wildcard Pharmaceutical Consulting. I had not anticipated the interest and number of readers this niche blog have attracted over the years so I have kept the historic blog posts. The company was voluntarily closed in the end of 2018 after having existed for eight years and been my full time occupation for four years. There may thus be historical references scattered in the old blog posts. I have removed some of the less visited posts, but you are welcome to drop me a note if you are missing a particular page or use the internet archive.
Science Based
I try to keep the blog posts instructive, usable and as light and easy as possible without going into too many details. The details, prerequisites and scientific rigor I put in my papers and pre-prints. Here’s a list of some of the most recent and relevant to this blog.
-
De Novo Molecular Design by Combining Deep Autoencoder Recurrent Neural Networks with Generative Topographic Mapping Boris Sattarov, Igor I. Baskin, Dragos Horvath, Gilles Gérard Marcou, Esben Jannik Bjerrum, and Alexandre Varnek. Journal of Chemical Information and Modeling, 2019
- Improving Chemical Autoencoder Latent Space and Molecular De Novo Generation Diversity with Heteroencoders. Esben Bjerrum and Boris Sattarov Biomolecules 8.4 (2018): 131.
- HIGH IMPACT Preprint: SMILES Enumeration as Data Augmentation for Neural Network Modeling of Molecules, Esben Jannik Bjerrum, Arxiv.org, 2017
- pICalculax: Improved Prediction of Isoelectric Point for Modified Peptides, Esben Jannik Bjerrum, Jan Holst Jensen, Jakob Lind Tolborg, Journal of Chemical Information and Modeling, 2017
- DeepIEP: a Peptide Sequence Model of Isoelectric Point (IEP/pI) using Recurrent Neural Networks (RNNs), Esben Jannik Bjerrum, Arxiv.org, 2017
- Molecular Generation with Recurrent Neural Networks (RNNs), Esben Jannik Bjerrum, Richard Threlfall, Arxiv.org, 2017
- Machine learning optimization of cross docking accuracy, Esben Jannik Bjerrum, Computational Biology and Chemistry, 2016
Guest Blogging
Now that the blog is not tied to my company as such, it could be interesting if others have something to write about that fits with the subjects. If you have something interesting to write about, such as showcasing your cool python chemistry toolkit, chemistry deep learning model or similar, please reach out.
Feel free to ask questions in the comments and happy deep cheminfo hacking
Esben Jannik Bjerrum