🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Thanks for trajectories!

Nice to have a wonderful lunch with Prof. Ken Dill from Stony Brook University to celebrate his winning of the Hirschfelder prize in theoretical chemistry. His passion about education and research inspired me a lot.TCI UW-Madison

Our MRS Bulletin article discussed perspectives in building kinetic network models to study molecular self-assembly! Congratulations to Bojun Liu Yunrui Qiu & Eshani! May thanks to @tiwarylab & the Kulik Group for the invitation to their special issue! rdcu.be/cX7Ao

Our preprint utilizes GraphVAMPnets to find CVs for self-assembly. Graph embeddings ensure structures invariant to permutations and rotations, and the VAMP theory (no detailed-balance requirement) handles insufficient sampling of dissociation transitions. chemrxiv.org/engage/chemrxi…

So happy to share the research progress with new friends and got recognized. Cannot make it without the help from my group members!

Our LPC method can efficiently group parallel kinetic pathways into distinct metastable path channels. It utilizes the variational autoencoder (VAE) to learn the spatial distributions of kinetic pathways and perform path clustering in the latent space. pubs.acs.org/doi/full/10.10…

Do the clustering of pathways according to their kinetic distribution using VAE !

Our IGME paper is out today! It efficiently models biomolecular dynamics by employing time integrations of memory kernels and avoiding numerical instability in time-dependent memory kernels. Thanks to The Journal of Chemical Physics for highlighting it as a Featured Article! pubs.aip.org/aip/jcp/articl…

OpenMM 8: Molecular Dynamics Simulation with Machine Learning Potentials. (arXiv:2310.03121v1 [physics.chem-ph]) #machinelearning #compchem arxiv.org/abs/2310.03121

Our non-Markovian dynamics models effectively reveal ensembles of metastable PPIs for PROTAC design. One of our findings matches the co-crystal structure perfectly, whereas AlphaFold3 struggles to predict these PPIs: None of its 50 predictions are suitable for PROTAC discovery.

It has been a truly rewarding and enjoyable experience working in the Huang Group over past four years. I am deeply grateful to Prof. Huang for the continuous help, support and encouragement. Will carry forward the spirit of collaboration and exploration into my future endeavors!

Our work on PROTAC has just been published in JACS Au! Congratulations to all the authors, especially Yunrui Qiu for driving this project! We use non-Markovian dynamic models to predict PPIs between KRAS & VHL for novel PROTAC design. pubs.acs.org/doi/10.1021/ja…