New CAR-T combinatorics cytotoxicity/stemness prediction

[kexinhuang12345]

Describe the problem
CAR-T is a promising therapeutics type that engineers T-cell receptors for cancer. The costimulatory domains of CAR is crucial for the output of T cells. Thus, a synthetic design of the domains would be critical for a better T cells throughput. There are lots of variabilities in the combinatorics designs characterized by different motifs. The machine learning question is given a combination (up to 3) of motifs, we want to predict the CAR phenotypes such as cytotoxicity. This problem is formulated by this great recent paper: https://www.biorxiv.org/content/10.1101/2022.01.04.474985v2.full.pdf

Describe the solution you'd like

from tdc.multi_pred import CART_Combo
data = CART_Combo(name = 'Daniels')

Additional context
Currently, the paper is still in preprint status, and the data is not open-sourced yet.