GPCR-PEnDB

GPCR-PEnDB provides a curated MySQL database of confirmed GPCR and non-GPCR protein sequences and associated sequence features to support prediction and classification of G protein-coupled receptors from amino acid sequences.

Key Features:

• Database: A searchable MySQL database compiling confirmed GPCR and non-GPCR protein sequences.
• Content and provenance: Includes 3,129 confirmed GPCR sequences and 3,575 non-GPCR sequences sourced from UniProtKB/Swiss-Prot representing over 1,200 species.
• Entry annotations: Provides source organism, classification, sequence length, composition, and derived sequence features for each protein entry.
• Datasets for methods: Supplies training and testing datasets for different combinations of computational tools, including machine learning and statistical methods.
• Comparative evaluation support: Enables assessment and comparison of multiple GPCR prediction tools.

Scientific Applications:

• Development and evaluation of prediction methods: Supports development and evaluation of computational GPCR prediction and classification methods.
• Machine learning and statistical modeling: Provides reliable training and testing datasets for machine learning and statistical approaches to GPCR prediction.
• Benchmarking: Enables comparative benchmarking of GPCR prediction tools across datasets.
• Biological research: Supports experimental and computational studies of GPCR function and classification across diverse organisms.

Methodology:

Compilation of sequences from UniProtKB/Swiss-Prot into a searchable MySQL database; annotation of each entry with source organism, classification, sequence length, composition, and derived sequence features; provision of training and testing datasets for combinations of computational tools to enable comparative performance assessment.