CareV: a contrastive learning framework for adaptive V gene representations.

This database contains processed data associated with the publication "ClareV: a contrastive learning framework for adaptive V gene representations." The archive includes two files:

Emerson_repertoire_raw.tar.gz: This file contains repertoire data from the Emerson cohort. The data have been processed through quality control, filtering, and full-length TCR assembly. ClareV project uses these repertoire data for TCR bag construction.

processed.tar.gz: This archive includes the processed data used in the publication. The tar.gz file includes two types of data packages: a V family-level bag and a V gene-level bag. Specifically:

superbags.pk: A TCR bag file split based on the IMGT-defined classification without TCR sampling (divided by V gene or family classification). 

weights.pk: Contains the corresponding frequency values of TCR clonotypes in each repertoire.

v_gene_order.pk: Records the order of V genes in both superbags.pk and weights.pk.

smp_list.pk and smp_labels.pk: Record the sample names and labels for the repertoire data in the TCR bags.