Introduction
G protein coupled receptors (GPCRs) represent a large family of membrane proteins, distinguished by their seven-transmembrane helical structures. These receptors play a pivotal role in numerous physiological processes. Nowadays, many researchers have proposed computational methods aimed at the identification of GPCRs. In the past, we introduced a powerful method, EMCBOW-GPCR, designed for this purpose. However, the feature extraction technique employed by it is susceptible to the out of vocabulary error, hinting potential for enhanced accuracy in GPCR identification. To solve the above-mentioned challenges, we propose a novel approach termed GPCR-AFPN. This method leverages the FastText algorithm to effectively extract features from protein sequences. Additionally, it employs a powerful deep neural network as the predictive model, to improve prediction accuracy. To validate the efficacy of the proposed GPCR-AFPN method, we conducted five-fold cross-validation and independent test respectively. The experimental results indicate that GPCR-AFPN outperforms existing methods.
Prediction by inputting sequences
The user can submit the protein sequences (example), and the webserver can predict whether the proteins entered by the user are GPCRs, and return the prediction result to the Web interface.
Prediction by inputting sequences
Users can submit a protein sequence file (FASTA format) by the function.The webserver will send the predicted results to the mailbox submitted by the user.
The explanation of Input:
Protein file: the FASTA format file containing the protein sequence to be predicted (example).
Program name: defined by users and the webserver will use it to identify the task.
Email: An email address to receive the prediction results after the prediction task is completed.
Open source
In addition to our online services, we extend our support to the academic community by providing access to the source code and models for download, thereby facilitating their research and application endeavors.