A map of all approved, illicit, and experimental drugs. A fully interactive Faerun visualization of @DrugBankDB embedded using TMAP from 2²³ dimensional MHFP space. This combines 4 projects I have been working on so far during my PhD. Check it out here:

.@PengkangG is at #ICML2025 with our paper on "Boosting Protein Graph Representations through Static-Dynamic Fusion". If you're around and want to step by:. Location: West Exhibition Hall B2-B3 hashtag#W-114 Time: Wednesday, July 16, 4:30 p.m. - 7 p.m. PDT

Join millions who have switched to Grok.

There's a couple of weeks left to apply 😊.

We're offering a fully funded PhD at the intersection of #ML/#AI and the life sciences with a focus on sustainability and chemistry. You'll work at WUR in NL, ranked #3 in environ. sciences, #1 in agricultural science, #38 in life sciences (QS). Apply:.

Thanks to the great work of @jirkalhotka, a former master's student at EPFL.

. and interpolate between conformations of a protein-DNA complex. Finally, we train an modulated autoencoder that learns the functions of over 500 small molecules and is able to also smoothly interpolate between them and embed them in a chemically meaningful latent space.

🧵 "Implicit Neural Representations of Molecular Vector-Valued Functions". New preprint in which we train modulated neural nets to approximate vector-valued functions representing molecules. We create superresolution repr. of low res protein-ligand pair.

If you're looking for a (fully funded) PhD position in #ml/#ai methodology (with plenty of opportunities to apply it) at the intersection of #chemistry and #biology, don't forget to apply before the end of the year.

Here's a first: I'm hiring a PhD student at @WUR. If you have (or know someone who has) a strong background and interest in computational (bio)chemistry or computer science (especially 3D computer vision), let's work together 😊.

. *However*, it also shows that the currently widespread benchmarks in machine learning for chemistry are a very low bar.

. It even be used on multimodal data, such as amino acid sequences and SMILES strings in binding affinity prediction tasks like this: s = gzip.encode(smiles + " " + aa) For regression or classification, we then use a weighted kNN and the normalised compression distance . .

In "Learning on compressed molecular representations" Jan Weinreich and I looked into whether GZIP performed better than Neural Networks in chemical ML tasks. Yes, you've read that right. TL;DR: Yes, GZIP can perform better than baseline GNNs and MLPs .

RT @fabiojcortes: Molecular dynamics simulations in mixed reality!. With @labriataphd and @lucien_krapp we’re scaling our multi-user WebX….

RT @simonduerr: Glad to share the first results from the ongoing user survey on BioIcons. Really appreciate that it's used around the globe….

. you will work with @CDB_77.

Great data science postdoc opportunity (French) in Grenoble:. Lethal Frontiers (FrontLet): Counting and Making Visible the Deaths at European Borders. This project is partnered with the UNITED campaign secretariat, Fatal Policies of Fortress Europe.

RT @difacquim: Happy to share a collaborative study generating and analyzing a fragment library of #naturalproducts from Colombia, Peru, an….

RT @reymondgroup: 🎊Our latest review "Chemical Space for Peptide-based Antimicrobials" by @MarkusOrsi, @HippolytePrsn, @bonvin15, @ThierryP….

RT @jjjvanderhooft: #CompMetabolomics #Annotation #MachineLearning.

RT @AGosztolai: Two weeks left to apply!.

RT @marianna_raps: 🔉 Our first BDSC Symposium on #AI in Clinical and Translational #Medicine is only 2 weeks away! . Are you an early caree….