chemspy.bsky.social

Preprint from the lab🚨
Have you ever engineered proteins to be more stable and were unhappy about your predictor's success rate? We got you covered with BoostMut!
Great work led by @kerlenkorbeld.bsky.social now online at www.biorxiv.org/content/10.1...
A thread 🧵

Inverse folding NNs are better at predicting changes to a mutated protein's equilibrium dynamics, relative to WT, than protein folding NNs (BioEmu & AF-Cluster). This was shown on an enzyme that is more enzymatically promiscuous in the open state.

De novo design of porphyrin-containing proteins as efficient and stereoselective catalysts De novo design of protein catalysts with high efficiency and stereoselectivity provides an attractive approach toward the design of environmentally benign catalysts. Here, we design proteins that inco...

Designing enzymes w/ a combination of structure-based sequence design (LigandMPNN) and directed evolution. The latter step discovers unconventional mutations like a mid-helical proline, and subsequent MD simulations show an increasingly rigidified active site

Systematic, computational discovery of multicomponent and one-pot reactions - Nature Communications Multi component reactions (MCRs) can build complex scaffolds from multiple starting materials in just one step without purification of intermediates but until now MCRs have been discovered rather than...

www.nature.com/articles/s41...

“A joint embedding of protein sequence and structure enables robust variant effect predictions”

Combines a structure GNN w/ MSA-transformer, partly by “only allowing attention between positions in the MSA that are proximal in” 3D space

www.biorxiv.org/content/10.1...
github.com/KULL-Centre/...

Really happy to share this iron photocatalyzed general alkene difunctionalization system developed by Kang-Jie, David, Xiaowei, Shih-Chieh, and James just published as a HOT article in Chemical Science!

pubs.rsc.org/en/content/a...

Altae-Tran, Kannan, Koonin, Zhang + al. reports @Science a new protein clustering algorithm + discovery of rare and previously unknown CRISPR-Cas systems - an untapped trove of diverse biochemical activities linked to RNA-guided mechanisms w/ potential for biotechnology development bit.ly/adi1910

“TRILL: Orchestrating Modular Deep-Learning Workflows for Democratized, Scalable Protein Analysis and Engineering” has been updated 🧪🧶🧬

Provides a framework to integrate PLMs, backbone diffusion, docking, inverse folding, etc

www.biorxiv.org/content/10.1...
trill.readthedocs.io/en/latest/ho...

Carbon-to-nitrogen single-atom transmutation of azaarenes - Nature A new type of transformation converting a heteroaromatic carbon atom into a nitrogen atom, turning quinolines into quinazolines to enable manipulation of molecular properties, is reported.

Today we report a new strategy for C-to-N replacement that leverages selective ring scission to turn a carbon atom into a leaving group while retaining the capacity for ring-reclosure around the nitrogen atom. Congrats to Jisoo, Colin, and Alec!

www.nature.com/articles/s41...

“Design of efficient artificial enzymes using crystallographically-enhanced conformational sampling” 🧪🧶🧬

Engineering a Kemp eliminase w/ 250-fold increase in enzymatic activity using an experimental structural ensemble and molecular dynamics-based filtering

www.biorxiv.org/content/10.1...

“The genetic architecture of protein stability” 🧪🧶🧬

An additive fitness prediction model trained on ≤2 mutations predicts fitness effects of 3-13 substitutions with R2=0.5

www.biorxiv.org/content/10.1...

Enzymatic synthesis and nanopore sequencing of 12-letter supernumerary DNA - Nature Communications Unnatural base pairing xenonucleic acids (XNAs) can be used to expand life’s alphabet beyond ATGC. Here, authors show strategies for enzymatic synthesis and next-generation nanopore sequencing of XN...

Here, we report on strategies for both writing and reading DNA with expanded alphabets composed of up to 12 letters (A, T, G, C, B, S, P, Z, X, K, J, V).

www.nature.com/articles/s41...

Fine-tuning protein Language Models by ranking protein fitness The self-supervised protein language models (pLMs) have demonstrated significant potential in predicting the impact of mutations on protein function and fitness, which is crucial for therapeutic...

The self-supervised protein language models (pLMs) have demonstrated significant potential in predicting the impact of mutations. Our method shows substantial improvements in the fitness prediction tasks even using less than 200 labeled data.

openreview.net/forum?id=DUj...

“OPUS-Rota5: A Highly Accurate Protein Side-chain Modeling Method with 3D-Unet and RotaFormer” 🧪🧶🧬

Better prediction quality than AlphaFold2 on CASP15 targets

www.biorxiv.org/content/10.1...
github.com/OPUS-MaLab/o...

“Predicting a Protein's Stability under a Million Mutations” 🧪🧶🧬

Structure-based stability prediction of single- and double-point mutants

arxiv.org/abs/2310.12979
github.com/jozhang97/Mu...

GitHub - sokrypton/af2bind Contribute to sokrypton/af2bind development by creating an account on GitHub.

Hmmm.....