Synthesis and characterization of ferrocene containing block copolymers

Narrowly dispersed diblock copolymers containing poly(methyl methacrylate) [PMMA] or poly(nonafluorohexyl methacrylate) [PF9MA] as the first block and poly(ferrocenylmethyl methacrylate) [PFMMA] as the second block, were prepared by anionic polymerization for the first time. Disordered bulk morphologies in the case of PMMA‐b‐PFMMA were observed and explained in terms of low Flory–Huggins interaction parameter (χ ≤ 0.04). In the case of PF9MA‐b‐PFMMA hexagonally packed cylinder morphology (HEX) was substantiated from TEM and SAXS observations. Furthermore, high incompatibility between PF9MA and PFMMA blocks allowed for the formation of well‐ordered ferrocene containing cylinders on silica substrate upon exposure of the thin films to a saturated solvent vapor. It was shown that the cylinder orientation, parallel or perpendicular to the surface, could easily be controlled by appropriate choice of the solvent and without the need for preliminary surface modification, for example by means of grafted brush layer. © 2016 Wiley Periodicals, Inc. J. Polym. Sci., Part A: Polym. Chem. 2017 , 55, 495–503     

FragNet,a Contrastive Learning-Based Transformer Model for Clustering,Interpreting, Visualizing,and Navigating Chemical Space

The question of molecular similarity is core in cheminformatics and is usually assessed via a pairwise comparison based on vectors of properties or molecular fingerprints. We recently exploited variational autoencoders to embed 6M molecules in a chemical space, such that their (Euclidean) distance within the latent space so formed could be assessed within the framework of the entire molecular set. However, the standard objective function used did not seek to manipulate the latent space so as to cluster the molecules based on any perceived similarity. Using a set of some 160,000 molecules of biological relevance, we here bring together three modern elements of deep learning to create a novel and disentangled latent space, viz transformers, contrastive learning, and an embedded autoencoder. The effective dimensionality of the latent space was varied such that clear separation of individual types of molecules could be observed within individual dimensions of the latent space. The capacity of the network was such that many dimensions were not populated at all. As before, we assessed the utility of the representation by comparing clozapine with its near neighbors, and we also did the same for various antibiotics related to flucloxacillin. Transformers, especially when as here coupled with contrastive learning, effectively provide one-shot learning and lead to a successful and disentangled representation of molecular latent spaces that at once uses the entire training set in their construction while allowing “similar” molecules to cluster together in an effective and interpretable way.