MORE | Spring 2024

Developing an Antibody Language Model for Generating Missing Amino Acid Residues to Complete Partial BCR Sequences

Health icon, disabled. A red heart with a cardiac rhythm running through it.

This research introduces a novel antibody language model for completing partial B cell receptor (BCR) sequences, addressing challenges in immune repertoire reconstruction from RNA-seq data. Leveraging AlphaFold’s EvoFormer embedding and attention mechanisms, the model learns contextual patterns within and across related BCR sequences. Trained on diverse antibody data, it aims to predict missing residues in partial sequences. Validation against TRUST4 on tumor RNA-seq data demonstrates its potential in capturing immune repertoire variability. This model contributes to efficiently profiling therapeutic immune cells, aiding cancer treatment development and infectious disease research.

Student researcher

Sonal Sujit Prabhu

Computer science

Hometown: Ankola, Karnataka, India

Graduation date: Spring 2025