Schematic representation of the learning approach of S2ALM Model (IMAGE)

Research

Schematic representation of the learning approach of S2ALM Model

Caption

Two-Stage Hierarchical Pre-training Framework of S2ALM: In Stage I, the model learns foundational sequence–structure relationships from large-scale protein data, using 1D amino acid sequences and 3D structural sequences. This stage employs masked language modeling to embed general biochemical patterns. Stage II shifts focus to antibody-specific data, incorporating Sequence–Structure Matching (SSM) and Cross-Level Reconstruction (CLR) objectives to capture the intricate interplay between antibody sequences and structures.

Credit

Copyright © 2025 Mingze Yin et al.

Usage Restrictions

Credit must be given to the creator.

License

CC BY

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.