Georg Seelig
@seeliglab
Very excited for MP3-seq, a new high-throughput Y2H approach we use to screen de novo protein heterodimer interactions. Fantastic work by Alex Baryshev, Alyssa La Fleur, @benjaminbgroves,@CirstynMichel, David Baker @UWproteindesign, @AjasjaLjubetic biorxiv.org/content/10.110…
biorxiv.org/content/10.110… Now on @biorxivpreprint find our paper describing how random sequences (N50) significantly affect splicing efficiency in yeast. We identified context specific secondary structure, motifs, and k-mers were most important in determining splicing, 1/
Benchmarks highlight GENA-LM's prowess in inferring diverse biological features from sequence data, from promoters & enhancers (data by @AlexanderStark8 lab) to splice sites & polyA signals (@seeliglab lab). It performs as well as, and sometimes even exceeds, existing models!5/n
Happy to share bioxriv preprint biorxiv.org/content/10.110… describing our new transformer models GENA-LM that can handle large genomic sequences up to 36 kb! Match or surpass previous models! Access on GH github.com/AIRI-Institute… & HF huggingface.co/AIRI-Institute #AI #Genomics 1/n
Stochasticity or memory that decides cell fate in daughter cells? "Both" seems to be the answer to this. They found "patterns" in daughter cell fates could explain many biologies. Lineage motifs: developmental modules for control of cell type proportions biorxiv.org/content/10.110…
It has been tremendously rewarding to work with outstanding @uwcse students, @HughChen18, @ianccovert, and @scottlundberg. Our paper that reviews and unifies 26 distinct algorithms to estimate Shapley values is just published in Nature MI! nature.com/articles/s4225…
Vector databases and similarity search are a cornerstone of ML systems, finally getting broad attention. Can't resist sharing some research results from a few years ago on using DNA-based molecular systems for very low energy large-scale similarity search. nature.com/articles/s4146…
.@SecGranholm has it wrong. We can’t greenlight the Mountain Valley Pipeline or any new fossil fuel project if we’re serious about moving to renewable energy. The idea that we’ll get to clean energy by tying ourselves to dirty energy for years to come is divorced from reality.
I am tremendously proud of my @UW_MSTP @uwcse student @joejanizek who developed an explainable AI approach for unsupervised gene expression modeling, which led to an exciting collaboration with @mkaeberlein's lab on understanding the Alzheimer's disease. genomebiology.biomedcentral.com/articles/10.11…
Excited to share our new paper published online at @nature! Congrats to the team, led by super talented students @JordanKesner1 and @chen_ziheng_cmu nature.com/articles/s4158…
Our latest study reveals an exciting new link between codon usage, cellular energy, and gene regulation. Ready for a short pitch? Go! biorxiv.org/content/10.110…
Excellent collaboration with @YongshengShi, Liang Liu, @jjohlin, and @seeliglab on sequence specificity of the anti-cancer compound JTE-607! Our ML model, Cleavage and Counteraction with Compound 2 on Polyadenylation Outcomes (C3PO), captures polyA sites' JTE-607 sensitivity. 🤖
New preprint from our lab and a great collaboration with @seeliglab ! The anti-cancer compound JTE-607 reveals hidden sequence specificity of the mRNA 3' processing machinery biorxiv.org/content/10.110…
Two new posts from @b_d_maier! One highlights a #preprint @seeliglab that reports high-throughput sequencing of protein-protein interactions, quantifying more than 100,000 interactions simultaneously. 👉 prelights.biologists.com/highlights/mas… Includes a summary of approaches to study PPIs!
New preprint from our lab and a great collaboration with @seeliglab ! The anti-cancer compound JTE-607 reveals hidden sequence specificity of the mRNA 3' processing machinery biorxiv.org/content/10.110…
To regulate cellular processes in #syntheticbiology it is useful to bring 2 proteins into proximity, to reconstitute function, guide localization, transcription... We report that from a single 4 helical bundle several pairs can be made. @NatureComms nature.com/articles/s4146… 🧵
Would you like to design de novo proteins and protein machines? 💻🔬🧬 There is an open PhD position in my group! Please RT and share! Application deadline is 15th of May.
My first😉first author paper just came out in Nature today🎉thank you to all the co-authors, collaborator Mike Bassik, and my co-advisors Polly @fordycelab & Lacra @BintuLab. Check out the supplement that contains plots of every protein's tiling results! tinyurl.com/CRTFNature
Care about the future of freedom and democracy in Wisconsin? Have a friendly dog? Help us Pet Out The Vote at noon tomorrow—Election Day!—on campus. mobilize.us/wisdems/event/…
The big banks are funding this massive new Alaska oil #WillowProject, even though they claim to worry about climate
The @ConocoPhillips #WillowProject would add to its already massive oil & gas expansion plans. This is enabled by the big US banks, which alone provided $11.6 billion to $COP in the years since the Paris Agreement. It's completely incompatible with banks' climate commitments.
We look towards a future of nanopore protein sequencing when single-molecule meets single-cell. scPeptide-barcoding, full-length protein reads, and “near-pore” processing…
A Comment by @KeisukeMotone and @jeffnivala highlights the challenges of fully decoding amino acids and their modifications using protein nanopore sequencing technology, and discusses its unique potential for single-cell proteomics. nature.com/articles/s4159…