This dataset is collected from TCRdb2.0.
The original sequence records are 691,744,135 with redundancy.
tcrdb2.txt removes duplicates and remains 290,314,598.
tcrdb2CDhit75.txt removes sequences with less than 75% identity using cd-hit and remains 1,782,927.
