If you wish to keep up with me, my mentoring, my science, and my community leadership, please follow me on LinkedIn (www.linkedin.com/in/jeffreygr...).
Posts by Jeffrey J. Gray
I’ve heard from friends who’ve suffered personal attacks with hate-filled content. I’ve worked to foster communities where all are welcome to share science and engineering to make the world better. TwiX is just not a safe place to share, so I will focus elsewhere.
I’m disgusted with X/Grok’s accelerating abusive use of sexualized images including even children. Kudos to Baltimore for fighting against this abuse. www.thebanner.com/politics-pow...
Over the last few months, I have experimented with a return to TwiX. But it has weighed on me. It’s just not a community environment consistent with my values. Therefore, I am again suspending my use of TwiX. A couple reasons recently:
Enjoying a swing through New England. After a fun time at @bpdmc.org, Ethan Eschbach hosted me for a visit with @sandercbio.bsky.social and the @deboramarks.bsky.social lab. I was also thrilled to give a talk at the Thayer School of Engineering at Dartmouth hosted by T. Emme Burgin. What a fun week!
Thank you @sokrypton.org and @cdbahl.com for your hospitality and congrats to you for creating a warm scientific community at @bpdmc.org. I hope our discussion sparked new ideas about accelerating protein design and antibody engineering. Grateful for the chance to share my lab’s work!
I'm so excited to speak at BPDMC this Wednesday evening! So many super protein scientists and modelers in Boston. I'm bringing some published and some unpublished work to share, hoping to spark some fun discussions.
Final stretch to apply for undergraduate summer internships in the Rosetta Commons! Come design proteins, develop AI and physics-based methods to model biomolecules, and impact health, materials, and sustainability! Application deadline is Sunday Feb 1.
rosettacommons.org/education/reu/
For the new year, I’ll start posting again on X. I’m frustrated by its algorithm, uncivil discourse, and the owner's politics. But our work can also reach a much broader audience. Grateful to Jeremy Wang helping manage my social media. Excited to share more of our work!
We updated our bioRxiv preprint last week to now include a succinct dimensional analysis of the major denoising diffusion equations. We summarize with a table that shows the meaning of each term in both AI and physics jargon.
Our goal in this project was to explore methods whereby one might extract some ‘potential’ from AlphaFold3 to see why it is so much better than our traditional physics based approaches.
As someone trained in physics, this project really helped me learn the nuances of denoising diffusion models. The “time” variable acts as an alchemical transformation—e.g. RFDiffusion transforms between protein structure and an ideal gas.
Read the paper here!
www.biorxiv.org/content/10.1...
Meet our new JHU ChemBE PhD student, Matthew Adrian! He studied ChemE at UC Berkeley and developed MLDL methods for small molecule drug property prediction at Merck. His research will improve MLDL for protein binding prediction! Fun fact: Matt is the starting goalie for JHU Ice Hockey! 🏒
@britnie-carpentier.bsky.social presenting her Rosetta energy approximation network at European RosettaCon!
In these pictures, we are at the historic castle of Bled…lab members Britnie, Natalia, Fatima and friend Rebecca are the antibody and I’m an antigen! Thanks to @ajasjaljubetic.bsky.social and Dominic Gront for being amazing conference chairs and hosts!
Somewhat belated, but here’s our lab contingent at European Rosetta Con in Slovenia! As always, the welcoming @rosettacommons.bsky.social provided incredible scientific discussions, collaborations, and networking. With @britnie-carpentier.bsky.social, Natalia Rincon, and @fatimolecule.bsky.social!
Congratulations to @swcanner.bsky.social for the successful defense of his thesis, “The Protein Sugar Interactome”! Leveraging deep learning and AI, he created tools and uncovered putative structural interactions underlying intercellular communication. He will receive his diploma in December.
A comparison of DFMDock performance to Rosetta's top ranked structure. Read the preprint here: www.biorxiv.org/content/10.1...
By using an additional prediction head for an energy with its gradient matched to the score, we extract an energy function and compare it to Rosetta, an empirically tuned physically-inspired energy function. The learned energy sometimes recovers funnels with near-native minima, and sometimes not!
Excited to share our new preprint “Unified Sampling and Ranking for Protein Docking with DFMDock” by PhD candidate Lee-Shin Chu and Post-Doc Sudeep Sarma! We designed a diffusion docking model that samples rigid-body conformations by predicting forces and ranks poses using predicted energies.
My latest manuscript is now on BioRxiv: Evaluation of De Novo Deep Learning Models on the Protein-Sugar Interactome. where we benchmark AF3 at protein-carbohydrate docking. This work was only possible with Dr. Lei Lu @takeshita-sho.bsky.social and @jeffreyjgray.bsky.social
doi.org/10.1101/2025...
Meet our new post-doc, Alejandra Montaño Romero! PhD in @cryoet.bsky.social’s lab (JHMI), she studied neurodevelopmental mutations in glutamate receptors via experimental & in silico work. Now she will improve deep learning for membrane proteins and design therapeutic antibodies targeting them!
Would love to follow jazzway here instead of... Anywhere else
Our post-bac program will give you a year of research training after you finish undergraduate. Work on incredible biomolecular structure and design projects (e.g. antibody engineering, or deep learning algorithms), and have great mentoring and support in applying for grad schools and fellowships.
Winter RosettaCon: Early registration deadline extended to Jan 8
Good news! The early registration deadline for Winter RosettaCon has been extended to January 8.
Additionally, the travel award deadline has been extended to January 6.
Learn more: winterrosettacon.org
Join us to research protein design! After that Nobel Prize, the field is wide open for impactful applications in medicine, sustainability, and technology. It's an incredible summer.
We're excited to work with @omsf.io OpenADMET and @polarishub.io to bring you this blind predictive modeling community challenge!
If you're developing predictive models for drug discovery---potency or ADMET---here's your chance to test them prospectively on real discovery data from ASAP!
Ice skating with the lab at the Johns Hopkins ice rink!
Deadline is Thurs Feb 1! Apply now for summer internships in computational biomolecular design.
(postbac apps due Feb 15)
Superfun at #NeurIPS to see lab alumni Jeliazko Jeliazkov and @dakpinaroglu.bsky.social, here with current PhD students Michael Chungyoun and Fatima Hitawala.