CoSiNE Training Dataset

Training data for CoSiNE model, which simulates antibody affinity maturation. Please read our paper for more details!

Dataset Summary

This dataset consists of approximately 2 million B-cell receptor (BCR) parent-child sequence transitions derived from 120,000 clonal families across 555 individual donors. The data was processed using a rigorous phylogenetic inference pipeline to capture the nuances of somatic hypermutation:

1. Clonal Inference: Sequences were clustered into clonal families and naive germlines were inferred using partis.
2. Quality Filtering: We retained only productive sequences (no stop codons, conserved CDR3 anchors) and excluded sequences with mutations in conserved signature cysteines.
3. Phylogenetic Reconstruction: Phylogenetic trees and ancestral sequences were inferred using IQ-TREE under a K80 substitution model.
4. Paired-Chain Modeling: For paired heavy and light chain data, we utilized an edge-linked-proportional partition model to account for distinct evolutionary rates across chains.

The final training set consists of Parent-Child Pairs (PCPs) extracted from the edges of these phylogenetic trees, representing a comprehensive map of the evolutionary trajectories within the adaptive immune system. For a detailed description of the processing pipeline, please refer to https://elifesciences.org/reviewed-preprints/109644v1.

Dataset Sources

The dataset was compiled using B-cell receptor (BCR) sequencing datasets from five sources:

• Jaffe-2022
• Tang-2022
• Vergani-2017
• Engelbrecht-2025
• Rodriguez-2023

Citation

Consider citing our paper if you use CoSiNE in your research!

@article{Lu2026ConditionallySN,
  title={Conditionally Site-Independent Neural Evolution of Antibody Sequences},
  author={Stephen Zhewen Lu and Aakarsh Vermani and Kohei Sanno and Jiarui Lu and IV FrederickA.Matsen and Milind Jagota and Yun S. Song},
  journal={ArXiv},
  year={2026},
  url={https://api.semanticscholar.org/CorpusID:285973749}
}

Dataset Card Contact

For questions or issues, please contact:

• Stephen Z. Lu (stephen.lu@berkeley.edu)
• Aakarsh Vermani (aakarshv@berkeley.edu)