DeepTMHMM
A transmembrane protein topology prediction model released by the Technical University of Denmark in 2022. It combines protein language model sequence embeddings with a hidden Markov model output layer to predict topology elements including transmembrane helices, transmembrane barrels, and signal peptides.
Timeline (last 90 days)
The research team used the same dataset employed during the development of the DeepTMHMM model and evaluated model performance using 5-fold cross-validation.
The paper does not disclose specific accuracy figures or a quantitative comparison with DeepTMHMM.
DeepTMHMM was released in 2022 by a team from the Technical University of Denmark and is a strong baseline model in the field of transmembrane protein topology prediction.
DeepTMHMM takes sequence embeddings generated by a protein language model as input and, combined with a structured output layer based on hidden Markov models, can predict multiple topology elements such as transmembrane helices, transmembrane barrels, and signal peptides.
All Facts (4)
The research team used the same dataset employed during the development of the DeepTMHMM model and evaluated model performance using 5-fold cross-validation.
50%UnverifiedThe paper does not disclose specific accuracy figures or a quantitative comparison with DeepTMHMM.
50%UnverifiedDeepTMHMM was released in 2022 by a team from the Technical University of Denmark and is a strong baseline model in the field of transmembrane protein topology prediction.
50%UnverifiedDeepTMHMM takes sequence embeddings generated by a protein language model as input and, combined with a structured output layer based on hidden Markov models, can predict multiple topology elements such as transmembrane helices, transmembrane barrels, and signal peptides.
50%