Congratulations to all the authors Aleksander Durumeric, Sean McCarty, Jay Smith, @jonkhler.argmin.xyz, @katarinaelez.bsky.social, LluΓs Raich and Patricia Suriana!!
These findings inspire a number of questions about how to translate fitness landscapes to prospective evolution predictions. Therefore, we introduce mavenets: github.com/SztainLab/ma... for reproducing and building upon this work.
MCMC simulations predict high fitness combinatorial variants with sequence profiles matching variants of concern, predicting omicron mutations using a model trained only near WT sequences.
We systematically evaluate performance on held out data within a single DMS experiment, data from independent DMS libraries, and epidemiological data, maintaining accuracy at edit distances much further from the training data, including omicron variants.
Top: absolute error of delta log KD of MLP, linear, sum, and mean of point mutants. The MLP remains low up to 9 mutants. Linear, sum, and mean are higher. Middle: Spearman's rank. MLP is highest, though others are close, all start to dip below 0.5 after 5 mutations. Bottom: distribution of mutations in sequences. Many contain 2-4, while less include 1 and 5-9.
These models predict the impact of combinatorial mutations more accurately than linearly combining the impact of point mutants, suggesting an understanding of epistatic relationships.
We combined machine learning, deep mutational scanning data, and Markov chain Monte Carlo simulations to characterize the SARS-CoV-2 RBD fitness landscape.
Graphical abstract "DMS binding data" above rendering of ACE2 bound SARS-CoV-2 spike RBD. "Machine learning" over neural network schematic. "Fitness landscape" schematic. "Predicted variants" showing sequence logo plot.
Preprint update!!!
How do mutations impact protein function? The answer is key to understanding evolving pathogens, predicting drug resistance, and designing custom proteins.
www.biorxiv.org/content/10.1101/2024.09.20.614179v2
Positions are open to both Assistant and Associate Professor ranks!
Picture of UM building
Using AI/ML for drug discovery? UM is hiring tenure track faculty at multiple ranks! please share and feel free to reach out if you have any questions.
careers.umich.edu/job_detail/2...
The AViDD centers were incredibly important for making scientific discoveries that improve human health. Hopefully they will be re-instated in some form
Grateful I was invited to share our recent efforts searching through sequence, structure, and chemical space to predict and combat antiviral drug resistance at the last Antiviral Drug Discovery (AViDD) Open Science Forum before it went on hold:https://youtu.be/FCZ0lFBs5sQ?si=Y7SBdksMSKI9Yedd
We have multiple ongoing projects related to protein engineering and antiviral drug resistance. Our lab uses a variety of dynamic computational and experimental techniques to unravel molecular mechanisms for rational design. Please send a CV and cover letter to be considered.
Currently interviewing candidates for Lab Manager Position
* expertise conducting high impact research with protein NMR spectroscopy required.
* this should be demonstrated by multiple first author publications.
Today Mortaza Derakhshani, PhD will be presenting "Enzyme Engineering for Sustainable Nylon Recycling" at the ACS Great Lakes Regional Meeting!
1) Free diffusion 2) electrostatic steering 3) encounter complex 4) energy barrier crossing 5) bound complex
PhD student in the lab, Daniel Santos Perez gave his first poster presentation today at the Midwest Theoretical Chemistry Conference!
Titled "Weighted ensemble enhanced sampling for the elucidation of drug binding pathways"
AlphaFold is amazing but gives you static structures π§
In a fantastic teamwork, @mcagiada.bsky.social and @emilthomasen.bsky.social developed AF2Ο to generate conformational ensembles representing side-chain dynamics using AF2 π
Code: github.com/KULL-Centre/...
Colab: github.com/matteo-cagia...
Robin, round bird with black/ grey top and orange tummy sitting on nest outside window of lab
Bird nest with three bright blue eggs
We have Easter eggs on the lab windowsill!!
Six people in lab coats standing in a circle listening to one explaining protocol
Six people in lab coats (and gloves and eye protection) at the bench with ice buckets and tubes
#Compchem experts can do experiments to! Everyone is learning how to test their protein mutation predictions this week π§ͺπ©βπ¬
This is the moment we've worked toward for a long time: First public disclosure of the @asapdiscovery.bsky.social pan-coronavirus antiviral aiming to help keep humanity safe from future pandemic threats like MERS-CoV and other bat coronaviruses.
Many of the sessions I've been in at #ACSSpring2025 have had almost exclusively (white) male speakers π Can folks please link some of your favorite speakers who could be invited to help balance this?
First they came for the Communists And I did not speak out Because I was not a Communist Then they came for the Socialists And I did not speak out Because I was not a Socialist Then they came for the trade unionists And I did not speak out Because I was not a trade unionist Then they came for the Jews And I did not speak out Because I was not a Jew Then they came for me And there was no one left To speak out for me Martin NiemΓΆller (1892 - 1984)
Hard to overhear some of these conversations of those who choose to be silent for fear of being targeted, when many of us dont have a choice of whether or not to be targeted.
Ombre curls pink to purple to blue
Desk with pink, purple, and blue pens
Similar colored figure of active learning cycle where a semi-emperical score function is used as a predictive model that is tested, for data to be fed in to refine the score function
I guess I have a color pallete
regarding Columbia University, I just learned about this (anonymous) account of negotiations of the Hitler regime in the 1930s with Frankfurt University -- Germany's most prestigious university at the time:
Who else will be in town? Would love to catch up with peeps :)
Drug Design: Machine Learning and Computational Methods Room 33A San Diego Convention Center Division/Committee: COMP: Division of Computers in Chemistry Organizers Kira Campbell, GlaxoSmithKline USA Olexandr Isayev, Carnegie Melon University Alexander Williams, GSK Presider Aaftaab Sethi, Uniwersytet im Adama Mickiewicza w Ponznaniu
4196324 - Physics-informed machine learning for computer-aided drug design 10:10 am - 10:25 am PDT Monday, March 24th, 2025 Room 33A San Diego Convention Center
Headed to San Diego βοΈ
for #ACSSpring2025
Resources for Federal Scientists from the Union of Concerned Scientists
Here are resources to have on hand during the Trump administration. We will continue to update this page as more resources are developed. π§ͺ
www.ucs.org/resources/re...
For instance, the immigration bar, and powerful Big Law pro bono practices, frequently coach clients to conceal their past or lie about their circumstances when asserting their asylum claims, all in an attempt to circumvent immigration policies enacted to protect our national security and deceive the immigration authorities and courts into granting them undeserved relief. Gathering
To address these concerns,| hereby direct the Attorney General to seek sanctions against attorneys and law firms who engage in frivolous, unreasonable, and vexatious litigation against the United States or in matters before executive departments and agencies of the United States.
Trump just directed the DOJ to attempt to disbar attorneys who take legal action against the executive branch while also indicating they will do the same to immigration attorneys.
www.whitehouse.gov/presidential...
It was great to catch up with friends at APS: Pilar Cassio, @rocovino.bsky.social Jutta Rogal, Berend Smit, Tess Smidt and @terrasztain.bsky.social
