SEMA (Spatial Epitope Modelling with Artificial intelligence) is a set of research tools for sequence- and structure-based conformational B-cell eptiope prediction, accurate identification of N-glycosylation sites, and a distinctive module for comparing the structures of antigen B-cell epitopes enhancing our ability to analyze and understand its immunogenic properties.
SEMA 2.0 contains following models:
SEMA-1D model is based on an ensemble of ESM2 transformer deep neural network protein language models.
SEMA-3D model is based on an ensemble of inverse folding models, SaProt.
The N-glycosylation prediction model (SEMA_PTM) was obtained by adding a fully-connected linear layer on the top layer of the ESM-2 pre-trained model.
Epitope comparison model is trained to identify local structural similarities within proteins, based on the non-linear transformation of multiplication of the embeddings of PLM with geometric modalities.