Protein Secondary Structure Prediction Using Transformers

Predicting protein secondary structures such as alpha helices, beta sheets, and coils from amino acid sequences is critical for understanding protein function. A transformer-based model is presented, applying attention mechanisms to protein sequence data for structural motif prediction. Data augmentation using a sliding window technique is employed on the CB513 dataset to augment the dataset. The transformer demonstrates strong potential in generalizing across variable-length sequences and capturing both local and long-range residue interactions.