*3.6. ADMET Profile*

In-silico ADMET properties of top-ranked hits were determined by deep learning models; more than 17 models were employed at the backend, which provided predictions on the ADME profile of each hit. It is an important part in drug development that can identify the desired pharmacological properties of compounds. In the current study, the message passing neural network (MPNN) is employed for the determination of ADMET properties. It was observed that compound **762** showed the lowest clinical toxicity value of 0.28%. The ADMET profile of top hits is tabulated in Table 9.

**Table 9.** ADMET properties of top hits predicted via MPNN model.

