Lin Wang (王林)

Lin Wang (王林)

A boy wandering in the beach of drug discovery.

Brief Introduction

I’m Lin Wang (王林), a Ph.D. student of ShanghaiTech University, supervised by Fang Bai (白芳). I received my B.S. degree at college of life science of North East Agricultural University (NEAU) in 2019.

My research interests are focused on developing artificial intelligence methods for lead discovery and applying them to design drugs for the treatment of cancer, inflammation, neurodegenerative diseases, and chronic illnesses. I have a particular interest in developing molecular modulators for protein conformational changes and protein-protein interactions. I am dedicated to advancing computational methods for molecular representation learning and molecular design to revolutionize drug discovery approaches.

Representative Works

  • GeminiMol: a generic molecular representation model pre-trained on a miniaturized dataset (39,290), which incorporates conformational space profile into molecular representation learning. BioRxiv GitHub
  • PPI-Miner: a motif-driven PPI identification method, that can be applied to discovering novel PPIs, the rational design of molecular glues, and protein vaccines. Using the PPI-Miner, we developed the first potential substrates database of CRBN, serving the rational design of molecular glue degraders. [Journal of Chemical Information and Modeling] [Online Database]
  • Discovery of PPI disruptors by targeting the SARS-CoV-2 spike protein. [Journal of Medicinal Chemistry | Acta Pharmacologica Sinica]

News

  1. 12/2023, we have introduced the GeminiMol, which incorporates conformational space information into molecular representation learning, refer to GitHub and BioRxiv.
  2. 10/2023, DeepSA was published on Journal of Cheminformatics, you can try it on our DeepSA webserver.
  3. 01/2023, we developed a highly sensitive model (DeepSA) for assessing the synthetic accessibility of small molecules, this work was done jointly by Shihang Wang (王世航) and me.
  4. 12/2022, we have upgraded the original plmd script and renamed it to AutoMD, which supports more force fields and provides the ability to customize the trajectory analysis pipeline.
  5. 11/2022, we developed a motif driven PPI mining method, named PPI-Miner. This methodology paper was published on Journal of Chemical Information and Modeling, and source code was stored in GitHub.
  6. CRBN is the most popular target for molecular glue and PROTAC protein degraders. In 08/2022, we released a database of potential substrates for CRBN, which covers all human proteins that contain the β-hairpin loop and complementary surface with CRBN.
  7. 08/2021, we found an allosteric drug binding pocket on Spike protein and designed allosteric inhibitors to prohibiting the conformational change of the Spike protein. This work was published on Journal of Medicinal Chemistry.
  8. 08/2021, we identified five potential small molecular anti-virus blockers via targeting SARS-CoV-2 spike protein and published this study on Acta Pharmacologica Sinica.
  9. 05/2021, I released a series of automated scripts for cross-docking (XDock), virtual screening (GVSrun), and molecular dynamics simulations (plmd) using the Schrödinger package, and stored the source code on GitHub.
  10. 05/2020, I build an user-friendly biomedical database-integration website for accelerating literature research for drug development researchers.
  11. 12/2019, I developed an automated script for downloading the single-chain protein structures by Uniprot ID, named GetPDB.