English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Figuring out small molecule binding sites in target proteins, in theresolution of either pocket or residue, is critical in many virtual and realdrug-discovery scenarios. Since it is not always easy to find such bindingsites based on domain knowledge or traditional methods, different deep learningmethods that predict binding sites out of protein structures have beendeveloped in recent years. Here we present a new such deep learning algorithm,that significantly outperformed all state-of-the-art baselines in terms of theboth resolutions$\unicode{x2013}$pocket and residue. This good performance wasalso demonstrated in a case study involving the protein human serum albumin andits binding sites. Our algorithm included new ideas both in the modelarchitecture and in the training method. For the model architecture, itincorporated SE(3)-invariant geometric self-attention layers that operate ontop of residue-level CNN outputs. This residue-level processing of the modelallowed a transfer learning between the two resolutions, which turned out tosignificantly improve the binding pocket prediction. Moreover, we developednovel augmentation method based on protein homology, which prevented our modelfrom over-fitting. Overall, we believe that our contribution to the literatureis twofold. First, we provided a new computational method for binding siteprediction that is relevant to real-world applications, as shown by the goodperformance on different benchmarks and case study. Second, the novel ideas inour method$\unicode{x2013}$the model architecture, transfer learning and thehomology augmentation$\unicode{x2013}$would serve as useful components infuture works.

Boosting Convolutional Neural Networks' Protein Binding Site Prediction  Capacity Using SE(3)-invariant transformers, Transfer Learning and  Homology-based Augmentation