Uneven TCR chain pairing constraints govern epitope recognition

This is processed TCR repertoire data from exo-TCR chain replacement experiments. a) clonotype_tables.zip contains the following folders: bulk: These are bulk TCRalpha and TCRbeta repertoires from CD4 and CD8 T cells of healthy PBMC donors before the exo-TCR experiments.For example:vdj.A12_CD4a_alpha.clonotypes.TRA.txt is the TCRalpha repertoire from CD4+ T cells of donor A12. epitope_specific: Contains subfolders corresponding to a given exo-TCR chain (e.g., A01_LTDb contains the result for introducing the A01_LTD epitope-specific TCRbeta chain), and each subfile contains repertoires of partner chains binding (pos) or non-binding (neg) to this epitope. (e.g., A01_LTDb/vdj.A12_LTDb_pos_alpha.clonotypes.TRA.txt contains data for binding TCRalpha chains in donor A12 after introducing the A01_LTD-specific TCRbeta chain). b) pos_neg.zip contains tab-separated text files with high confidence binding and non-binding TCRalpha and TCRbeta clonotypes for each epitope after applying all filters to data from epitope_specific folder and aggregating across donors (see manuscript for details). Each data table in this folder is a merge between positive and negative, resulting in duplication of column names with .x and .y suffices. For example, in EIYb_true_pos.tsv, Read.count.x is UMI count of given clone in positive (epitope binding) and Read.count.y is UMI count in negative (non-binding). For EIYb_true_neg.tsv this order is inverted, Read.count.x is UMI count of given clone in negative, Read.count.y is UMI count in positive.c) close_in_seq_pos_neg.zip contains .csv files with a subset of sequences from pos_neg folder above for GIL and LLW epitopes. LLWa_close_negatives.csv and GILb_close_negatives.csv contain non-binding (negative) sequences with a single amino acid mismatch neighbour among binding (positive sequences), LLWa_close_positive.csv and GILb_close_positive.csv contain binding sequences witha single amino acid mismatch among negative sequences.