suwenbin.bsky.social
@suwenbin.bsky.social
Reposted by suwenbin.bsky.social
New preprint with @mcagiada.bsky.social & @sokrypton.org in which we present a benchmark and predictions of absolute protein stability (ΔG not ΔΔG) using using likelihoods from a generative model, and also benchmark it for conformational free energies against NMR 🧬 🧶
doi.org/10.1101/2024...
doi.org/10.1101/2024...
March 16, 2024 at 10:21 AM
New preprint with @mcagiada.bsky.social & @sokrypton.org in which we present a benchmark and predictions of absolute protein stability (ΔG not ΔΔG) using using likelihoods from a generative model, and also benchmark it for conformational free energies against NMR 🧬 🧶
doi.org/10.1101/2024...
doi.org/10.1101/2024...
Reposted by suwenbin.bsky.social
PLMs trained on sequence/structural data that is highly local (e.g., for a single protein family) outperform those trained on global datasets (e.g., all protein sequences)
doi.org/10.1101/2024.03.15.585128
doi.org/10.1101/2024.03.15.585128
March 17, 2024 at 9:30 PM
PLMs trained on sequence/structural data that is highly local (e.g., for a single protein family) outperform those trained on global datasets (e.g., all protein sequences)
doi.org/10.1101/2024.03.15.585128
doi.org/10.1101/2024.03.15.585128
Reposted by suwenbin.bsky.social
Predicting absolute protein folding stability using generative models https://www.biorxiv.org/content/10.1101/2024.03.14.584940v1
Predicting absolute protein folding stability using generative models https://www.biorxiv.org/content/10.1101/2024.03.14.584940v1
While there has been substantial progress in our ability to predict changes in protein stability due
www.biorxiv.org
March 16, 2024 at 12:48 AM
Predicting absolute protein folding stability using generative models https://www.biorxiv.org/content/10.1101/2024.03.14.584940v1
Reposted by suwenbin.bsky.social
The more worrisome outcome is that tech companies endanger patients by strong-arming their AI-designed de novo proteins into clinical trials without appropriate in vivo data, akin to what happened with self-driving cars/social media algorithms/a million other things
www.nature.com/articles/d41...
www.nature.com/articles/d41...
Could AI-designed proteins be weaponized? Scientists lay out safety guidelines
AI tools that can come up with protein structures at the push of a button should be used safely and ethically, say researchers in the field. AI tools that can come up with protein structures at the pu...
www.nature.com
March 13, 2024 at 7:50 AM
The more worrisome outcome is that tech companies endanger patients by strong-arming their AI-designed de novo proteins into clinical trials without appropriate in vivo data, akin to what happened with self-driving cars/social media algorithms/a million other things
www.nature.com/articles/d41...
www.nature.com/articles/d41...
Reposted by suwenbin.bsky.social
“ProSTAGE: Predicting Effects of Mutations on Protein Stability by Using Protein Embeddings and Graph Convolutional Networks"
Obtains “high predictive accuracy even when using AlphaFold2 predicted structures as input”
doi.org/10.1021/acs....
github.com/GenScript-IB...
Obtains “high predictive accuracy even when using AlphaFold2 predicted structures as input”
doi.org/10.1021/acs....
github.com/GenScript-IB...
January 23, 2024 at 6:10 AM
“ProSTAGE: Predicting Effects of Mutations on Protein Stability by Using Protein Embeddings and Graph Convolutional Networks"
Obtains “high predictive accuracy even when using AlphaFold2 predicted structures as input”
doi.org/10.1021/acs....
github.com/GenScript-IB...
Obtains “high predictive accuracy even when using AlphaFold2 predicted structures as input”
doi.org/10.1021/acs....
github.com/GenScript-IB...
Reposted by suwenbin.bsky.social
“Tpgen: a language model for stable protein design with a specific topology structure” has been published
doi.org/10.1186/s12859-024-05637-5
doi.org/10.1186/s12859-024-05637-5
January 24, 2024 at 6:41 AM
“Tpgen: a language model for stable protein design with a specific topology structure” has been published
doi.org/10.1186/s12859-024-05637-5
doi.org/10.1186/s12859-024-05637-5
Reposted by suwenbin.bsky.social
“Exploring Transition States of Protein Conformational Changes via Out-of-Distribution Detection in the Hyperspherical Latent Space”
Shows high energy states "exhibit a distributional shift from metastable states”
doi.org/10.26434/chemrxiv-2024-r8gjv
github.com/xuhuihuang/ts-dart
Shows high energy states "exhibit a distributional shift from metastable states”
doi.org/10.26434/chemrxiv-2024-r8gjv
github.com/xuhuihuang/ts-dart
January 24, 2024 at 6:40 AM
“Exploring Transition States of Protein Conformational Changes via Out-of-Distribution Detection in the Hyperspherical Latent Space”
Shows high energy states "exhibit a distributional shift from metastable states”
doi.org/10.26434/chemrxiv-2024-r8gjv
github.com/xuhuihuang/ts-dart
Shows high energy states "exhibit a distributional shift from metastable states”
doi.org/10.26434/chemrxiv-2024-r8gjv
github.com/xuhuihuang/ts-dart
Reposted by suwenbin.bsky.social
“TM-search: An Efficient and Effective Tool for Protein Structure Database Search” has been published
Applied TM-Align and iterative clustering to sift through huge DBs of protein structures faster than foldseek
doi.org/10.1021/acs....
Applied TM-Align and iterative clustering to sift through huge DBs of protein structures faster than foldseek
doi.org/10.1021/acs....
January 26, 2024 at 6:10 AM
“TM-search: An Efficient and Effective Tool for Protein Structure Database Search” has been published
Applied TM-Align and iterative clustering to sift through huge DBs of protein structures faster than foldseek
doi.org/10.1021/acs....
Applied TM-Align and iterative clustering to sift through huge DBs of protein structures faster than foldseek
doi.org/10.1021/acs....
Reposted by suwenbin.bsky.social
“HybridDBRpred: improved sequence-based prediction of DNA-binding amino acids using annotations from structured complexes and disordered proteins” has been published
doi.org/10.1093/nar/...
doi.org/10.1093/nar/...
January 26, 2024 at 6:10 AM
“HybridDBRpred: improved sequence-based prediction of DNA-binding amino acids using annotations from structured complexes and disordered proteins” has been published
doi.org/10.1093/nar/...
doi.org/10.1093/nar/...
Reposted by suwenbin.bsky.social
“TransMEP: Transfer learning on large protein language models to predict mutation effects of proteins from a small known dataset”
www.biorxiv.org/content/10.1...
github.com/strodel-grou...
www.biorxiv.org/content/10.1...
github.com/strodel-grou...
TransMEP: Transfer learning on large protein language models to predict mutation effects of proteins...
bioRxiv - the preprint server for biology, operated by Cold Spring Harbor Laboratory, a research and educational institution
www.biorxiv.org
January 16, 2024 at 7:01 AM
“TransMEP: Transfer learning on large protein language models to predict mutation effects of proteins from a small known dataset”
www.biorxiv.org/content/10.1...
github.com/strodel-grou...
www.biorxiv.org/content/10.1...
github.com/strodel-grou...
Reposted by suwenbin.bsky.social
“TemStaPro: protein thermostability prediction using sequence representations from protein language models” has been updated
www.biorxiv.org/content/10.1...
github.com/ievapudz/Tem...
www.biorxiv.org/content/10.1...
github.com/ievapudz/Tem...
January 17, 2024 at 6:29 AM
“TemStaPro: protein thermostability prediction using sequence representations from protein language models” has been updated
www.biorxiv.org/content/10.1...
github.com/ievapudz/Tem...
www.biorxiv.org/content/10.1...
github.com/ievapudz/Tem...
Reposted by suwenbin.bsky.social
“ProteinMPNN Recovers Complex Sequence Properties of Transmembrane β-Barrels”
Inverse folding methods implicitly learn membrane protein energy functions
www.biorxiv.org/content/10.1...
github.com/marissadolor...
Inverse folding methods implicitly learn membrane protein energy functions
www.biorxiv.org/content/10.1...
github.com/marissadolor...
January 18, 2024 at 5:33 AM
“ProteinMPNN Recovers Complex Sequence Properties of Transmembrane β-Barrels”
Inverse folding methods implicitly learn membrane protein energy functions
www.biorxiv.org/content/10.1...
github.com/marissadolor...
Inverse folding methods implicitly learn membrane protein energy functions
www.biorxiv.org/content/10.1...
github.com/marissadolor...
Reposted by suwenbin.bsky.social
“xTrimoPGLM: Unified 100B-Scale Pre-trained Transformer for Deciphering the Language of Protein” has been updated
This version describes xT-Fold, which appends the ESMFold trunk to the 100B PLM with favorable results
www.biorxiv.org/content/10.1...
This version describes xT-Fold, which appends the ESMFold trunk to the 100B PLM with favorable results
www.biorxiv.org/content/10.1...
January 12, 2024 at 6:23 AM
“xTrimoPGLM: Unified 100B-Scale Pre-trained Transformer for Deciphering the Language of Protein” has been updated
This version describes xT-Fold, which appends the ESMFold trunk to the 100B PLM with favorable results
www.biorxiv.org/content/10.1...
This version describes xT-Fold, which appends the ESMFold trunk to the 100B PLM with favorable results
www.biorxiv.org/content/10.1...
Reposted by suwenbin.bsky.social
“OmeDDG: Improved Protein Mutation Stability Prediction Based on Predicted 3D Structures” has been published
Stability prediction using OmegaFold models
doi.org/10.1021/acs....
Stability prediction using OmegaFold models
doi.org/10.1021/acs....
January 12, 2024 at 6:23 AM
“OmeDDG: Improved Protein Mutation Stability Prediction Based on Predicted 3D Structures” has been published
Stability prediction using OmegaFold models
doi.org/10.1021/acs....
Stability prediction using OmegaFold models
doi.org/10.1021/acs....
Reposted by suwenbin.bsky.social
“Tree Search-Based Evolutionary Bandits for Protein Sequence Optimization”
A bandit-based sequence design approach for viral capsids
arxiv.org/abs/2401.06173
A bandit-based sequence design approach for viral capsids
arxiv.org/abs/2401.06173
January 16, 2024 at 7:00 AM
“Tree Search-Based Evolutionary Bandits for Protein Sequence Optimization”
A bandit-based sequence design approach for viral capsids
arxiv.org/abs/2401.06173
A bandit-based sequence design approach for viral capsids
arxiv.org/abs/2401.06173
Reposted by suwenbin.bsky.social
“PRO-LDM: Protein Sequence Generation with a Conditional Latent Diffusion Model” has been updated
www.biorxiv.org/content/10.1...
www.biorxiv.org/content/10.1...
January 16, 2024 at 7:01 AM
“PRO-LDM: Protein Sequence Generation with a Conditional Latent Diffusion Model” has been updated
www.biorxiv.org/content/10.1...
www.biorxiv.org/content/10.1...
Reposted by suwenbin.bsky.social
“Ancestral Sequence Reconstruction as a tool to detect and study de novo gene emergence”
Authors describe a workflow for de novo gene identification and find twenty genes that they believe emerged after the speciation of S. cerivisiae
www.biorxiv.org/content/10.1...
github.com/Nikos22/huma...
Authors describe a workflow for de novo gene identification and find twenty genes that they believe emerged after the speciation of S. cerivisiae
www.biorxiv.org/content/10.1...
github.com/Nikos22/huma...
January 3, 2024 at 6:50 AM
“Ancestral Sequence Reconstruction as a tool to detect and study de novo gene emergence”
Authors describe a workflow for de novo gene identification and find twenty genes that they believe emerged after the speciation of S. cerivisiae
www.biorxiv.org/content/10.1...
github.com/Nikos22/huma...
Authors describe a workflow for de novo gene identification and find twenty genes that they believe emerged after the speciation of S. cerivisiae
www.biorxiv.org/content/10.1...
github.com/Nikos22/huma...
Reposted by suwenbin.bsky.social
“ThermoFinder: A sequence-based thermophilic proteins prediction framework”
www.biorxiv.org/content/10.1...
github.com/Luo-SynBioLa...
www.biorxiv.org/content/10.1...
github.com/Luo-SynBioLa...
January 3, 2024 at 6:50 AM
“ThermoFinder: A sequence-based thermophilic proteins prediction framework”
www.biorxiv.org/content/10.1...
github.com/Luo-SynBioLa...
www.biorxiv.org/content/10.1...
github.com/Luo-SynBioLa...
Reposted by suwenbin.bsky.social
“Protein language model powers accurate and fast sequence search for remote homology” has been updated
One of several recent methods to use embeddings rather than AA ID for sequence search/alignment problems
www.biorxiv.org/content/10.1...
dmiip.sjtu.edu.cn/PLMSearch
One of several recent methods to use embeddings rather than AA ID for sequence search/alignment problems
www.biorxiv.org/content/10.1...
dmiip.sjtu.edu.cn/PLMSearch
January 5, 2024 at 6:42 AM
“Protein language model powers accurate and fast sequence search for remote homology” has been updated
One of several recent methods to use embeddings rather than AA ID for sequence search/alignment problems
www.biorxiv.org/content/10.1...
dmiip.sjtu.edu.cn/PLMSearch
One of several recent methods to use embeddings rather than AA ID for sequence search/alignment problems
www.biorxiv.org/content/10.1...
dmiip.sjtu.edu.cn/PLMSearch
Reposted by suwenbin.bsky.social
“State-of-the-RNArt: benchmarking current methods for RNA 3D structure prediction” has been updated
A review of how different methods have evolved over the last ~15 years
www.biorxiv.org/content/10.1...
A review of how different methods have evolved over the last ~15 years
www.biorxiv.org/content/10.1...
January 6, 2024 at 7:51 AM
“State-of-the-RNArt: benchmarking current methods for RNA 3D structure prediction” has been updated
A review of how different methods have evolved over the last ~15 years
www.biorxiv.org/content/10.1...
A review of how different methods have evolved over the last ~15 years
www.biorxiv.org/content/10.1...
Reposted by suwenbin.bsky.social
“Homology detection using a protein secondary structure-based large language model”
www.biorxiv.org/content/10.1...
github.com/soroushv-dar...
www.biorxiv.org/content/10.1...
github.com/soroushv-dar...
December 21, 2023 at 4:36 AM
“Homology detection using a protein secondary structure-based large language model”
www.biorxiv.org/content/10.1...
github.com/soroushv-dar...
www.biorxiv.org/content/10.1...
github.com/soroushv-dar...
Reposted by suwenbin.bsky.social
“Position-specific evolution in transcription factor binding sites, and a fast likelihood calculation for the F81 model” has been updated
Uses “position-specific stationary vectors” like PSSMs for nucleotides instead of AAs
www.biorxiv.org/content/10.1...
github.com/rsidd120/TFB...
Uses “position-specific stationary vectors” like PSSMs for nucleotides instead of AAs
www.biorxiv.org/content/10.1...
github.com/rsidd120/TFB...
Position-specific evolution in transcription factor binding sites, and a fast likelihood calculation...
bioRxiv - the preprint server for biology, operated by Cold Spring Harbor Laboratory, a research and educational institution
www.biorxiv.org
December 22, 2023 at 10:31 AM
“Position-specific evolution in transcription factor binding sites, and a fast likelihood calculation for the F81 model” has been updated
Uses “position-specific stationary vectors” like PSSMs for nucleotides instead of AAs
www.biorxiv.org/content/10.1...
github.com/rsidd120/TFB...
Uses “position-specific stationary vectors” like PSSMs for nucleotides instead of AAs
www.biorxiv.org/content/10.1...
github.com/rsidd120/TFB...
Reposted by suwenbin.bsky.social
“Leveraging ancestral sequence reconstruction for protein representation learning”
A “small but focused” PLM for specific families ends up yielding more accurate sequences and smoother fitness landscapes than ESM-1b
www.biorxiv.org/content/10.1...
A “small but focused” PLM for specific families ends up yielding more accurate sequences and smoother fitness landscapes than ESM-1b
www.biorxiv.org/content/10.1...
December 22, 2023 at 10:32 AM
“Leveraging ancestral sequence reconstruction for protein representation learning”
A “small but focused” PLM for specific families ends up yielding more accurate sequences and smoother fitness landscapes than ESM-1b
www.biorxiv.org/content/10.1...
A “small but focused” PLM for specific families ends up yielding more accurate sequences and smoother fitness landscapes than ESM-1b
www.biorxiv.org/content/10.1...
Reposted by suwenbin.bsky.social
“Prot2Text: Multimodal Protein's Function Generation with GNNs and Transformers” has been updated
arxiv.org/abs/2307.14367
github.com/hadi-abdine/...
arxiv.org/abs/2307.14367
github.com/hadi-abdine/...
December 23, 2023 at 4:40 AM
“Prot2Text: Multimodal Protein's Function Generation with GNNs and Transformers” has been updated
arxiv.org/abs/2307.14367
github.com/hadi-abdine/...
arxiv.org/abs/2307.14367
github.com/hadi-abdine/...
Reposted by suwenbin.bsky.social
“ESM-Scan - a tool to guide amino acid substitutions”
www.biorxiv.org/content/10.1...
huggingface.co/spaces/thaid...
gitlab.tugraz.at/D5B8E3502557...
www.biorxiv.org/content/10.1...
huggingface.co/spaces/thaid...
gitlab.tugraz.at/D5B8E3502557...
December 13, 2023 at 6:06 AM
“ESM-Scan - a tool to guide amino acid substitutions”
www.biorxiv.org/content/10.1...
huggingface.co/spaces/thaid...
gitlab.tugraz.at/D5B8E3502557...
www.biorxiv.org/content/10.1...
huggingface.co/spaces/thaid...
gitlab.tugraz.at/D5B8E3502557...