.png){kind=icon}
Return
DeepAlloWeb: A Web Server for Interactive Allosteric Pockets Prediction Using Protein Language Model
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jmb.2026.169863.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• DeepAlloWeb may support drug discovery research by accurately predicting allosteric pockets, which in turn could help targeted therapy developments. • The webserver integrates a fine-tuned protein language model (ProtBERT) with multitask learning on allosteric pockets data. • Users can visualize the pLM attention among residues, which can provide information as to which residues are more important for a certain allosteric pocket residue for a certain layer and head of the pLM.
Keywords:
DeepAlloWeb
allosteric pockets
protein language model
drug discovery
multitask learning
Journal
IF:
4.5
Papers:
657
Citations:
5.4W
Organization
Cited Papers
ShareSaveAnalyzing the Structure of Attention in a Transformer Language Model.png)
Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP
OAAI
Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLPIF0
OAAIShareSave
ShareSaveShareSaveShareSave
View more