Yang Shen

ST-PARM: Pareto-Complete Inference-Time Alignment for Multi-Objective Protein Design biorxiv.org/content/10.648… #biorxiv_bioinfo

Congratulations to Yihong for his first conference presentation, just three months into his Ph.D. program! 🎉 Thank you CAGI7 @CAGInews and @NIH for sponsoring the travel fellowship and nurturing the next generation of scientific trainees. #CAGI #NIH #Genomics #VariantEffect #AI

Today CZI is announcing an unprecedented new scientific initiative to build the future of AI-powered biology. I am joining CZI to lead this initiative as Head of Science, and the EvolutionaryScale team is joining forces with Biohub. This is the first large scale scientific effort to combine frontier AI and frontier biology. I feel an incredible sense of optimism in this moment. There is a path to build predictive models of life that can fundamentally accelerate science, and unlock a new understanding of disease. biohub.org/blog/frontier-…

We are #hiring for Postdoctoral Associate in AI4Science at Texas A&M University. Message me if you're interested in joining our team. We are attending Conference on Neural Information Processing Systems (NeurIPS 2024) if you would like to meet! #NeurIPS2024

This paper was a collaborative effort from @marinkazitnik @_michellemli @aydin_wells @GysiDeisy @compbiologist @t_m_murali @pedjagogue @sroyyors @BaudotAnais @bozdags @kapil_devkota_ @anthonygitter @sargoshoe @piehiram @Meng_CS @ziynet_ @koyuturkm @jmuiuc @xanderpico @NatasaPrzulj @benjraphael @anna_m_ritz @StructBioinf @profmonasingh @dkslonim @haiyuanyu @ProfMilenkovic

Just published: "Current and future directions in network biology", a comprehensive review by leading experts in the field! Read it here: doi.org/10.1093/bioadv… #NetworkBiology #ISCB

.@bozdags, Byung-Jun Yoon, @structbioinf, & I led the section on Machine Learning on Networks w/ contributions from Ziynet Nesibe Kesimoglu & @haiyuanyu. Thanks to @marinkazitnik for tons of input.

Taking a Deep Dive into Chemical Space, a symposium at #ACSfall2023 biosolveit.de/2023/06/23/acs…

It’s great to be back at MPI after almost five years, see many familiar faces again, and hearing great science progress in modeling of protein interactions! #MPI2023

70 years ago on the day of Apr. 25, a short but revolutionary paper was published. nature.com/articles/17173… Let's keep in mind the historic contributions from Watson, Crick, Rosalind Franklin and others. Happy National DNA Day!

Finally, after ~6 years of work, this is published! genomebiology.biomedcentral.com/articles/10.11… Thanks to all my co-authors and the participants of the challenge for seeing this through.

@Stephen05485935 - Structure-informed Protein Language Models Wuwei Tan - Multimodal learning of noncoding variant effects using genome sequence and chromatin structure biorxiv.org/content/10.110…

Look forward to attending #bps2023 in just a few days! My students will present their works on predicting the effects of protein and noncoding variants. We hope to connect with more experimental researchers on #ProteinDesign!

It's frustrating reading comp bio articles these days because many keep falling into the same pitfalls. Hard to know if the method actually works, or whether they messed up the evaluation. Here are some issues I've seen recently (w/o names):

11 ways ChatGPT saves me hours of work every day, and why you'll never outcompete those who use AI effectively. A list for those who write code: 1 of 16

Multimodal learning of noncoding variant effects using genome sequence and chromatin structure biorxiv.org/cgi/content/sh… #biorxiv_bioinfo

In two new papers we have found that the ESM2 language model generalizes beyond natural proteins, and enables programmable generation of complex and modular protein structures.

Great pleasure to introduce a new study on single sequence protein structure prediction rdcu.be/c1Ubu and share my views rdcu.be/c1UbO ! @SpringerNature @NatComputSci

Great perspective on the field of protein structure prediction since #AlphaFold 2 and the winning solutions at #CASP15.

4. Embrace the model as part of the protocol. Most are using AF2 as a sampling tool now. One would think that scoring would be a coupled challenge. Yes and no.

3. Retrain/refine/customize the model. This was done by few groups for now and will be done by more in future. It needs clear rational / loss function, big and clean data, and BIG computing resources.

It is interesting to see how the community of protein structure prediction has responded to AlphaFold2 (AF2) (2020-21) while participating in the next competition of #CASP15. The short answer is: "if you can't beat them, join them". But there is more.